Rice hulls used as a starting material for the preparation of activated carbon were obtained from the Egyptian Starch & Glucose Manufacturing Company (ESGC).

The activated carbons from rice hulls (RHAC) were prepared according to the procedure described by [11] with some modification. Rice hulls were cleaned with water and dried at 110˚C for 48 h. The dried samples were crushed with a blender and sieved before they were carbonized at 400˚C for 1h in a muffle furnace in order to produce charcoal. The charcoals obtained were then subjected to impregnation in NaOH solution by weight ratio (1 g charcoal: 4 g NaOH) at 70˚C for 24 h. After that, it was dried in oven at 110˚C for 24 h. The resulted samples were further activated in a muffle furnace at 800˚C for 2 hrs. After cooling, the activated carbons were washed successively several times with 1 M HCl followed by hot water until the pH became neutral (=7). Finally the washed samples were dried at 110˚C.

The prepared activated carbons were characterized by Fourier transformation infrared technique (FTIR) over the range of 500 - 4000 cm⁻¹ using Thermo Nicolet Avatar 370 FTIR Spectrometer, Thermo scientific co. The surface characteristics of the adsorbents were also investigated by scanning electron microscope (SEM) using JEOL, JSM-6490LA Scanning Electron Microscope―JEOL USA, Inc.

Batch experiments were conducted to study the factors affecting the adsorption of Al(III) and Fe(III) onto RHAC. At the beginning of each experimental run, a known weight of RHAC was added to 50 ml solution containing a known concentration of single or binary aluminium(III) and iron(III) ions. The studied factors were pH (2 - 5), adsorbent dosage (0.5 - 2.0 g/l), initial ion concentration (5 - 100 mg/l), and contact time (30 - 240 min). The flasks were agitated in a shaking water bath at a 200 rpm constant shaking rate until equilibrium was reached. The mixture was then filtered and the remaining aluminium and/or iron concentrations were determined at 396.153 and 259.939 nm, respectively using atomic absorption spectrometer (Shimadzu, model AA-6300, Japan).

The amount of ions adsorbed at equilibrium (adsorption capacity), q_e (mg/g), was calculated by the following equation:

whereas the ions removal percentage (R%) was calculated by the following equation:

where C_i and C_e are the ions concentrations at the initial time and at equilibrium (mg/l), respectively. V is the volume of the solution (l) and W is the mass of adsorbent used (g).

Seven water samples (S1, S2, S3, S4, S5, S6 and S7; three replicates each) were collected from different locations as given in Table 1. Water sampling was achieved by “grab sampling technique” according to the “EPA Guidelines for water and wastewater sampling” [12] .

To compare the efficiency of RHAC and CAC for the removal of aluminium and iron from real wastewaters: 50 ml of each water sample were mixed with 0.1 g of RHAC or commercial activated carbon (CAC) obtained from Norit co. The samples were shaken for 240 min then they were separated from the adsorbent by filtration and finally the remaining concentrations of metal ions were determined.

The removal of Al(III) and Fe(III) from their single ion solutions as a function of contact time at different initial ions concentration is represented in Figure 1. It can be seen that the adsorption of both ions was low at the start, then increased by increasing the contact time and attained its maximum value after 180 min. of contact time with the studied adsorbent. The adsorption capacity of RHAC increased by increasing the initial ions concentration in the solution.

Figure 2 shows the removal percentages of Al(III) and Fe(III) in single and binary systems at various contact time and at an initial ions concentration of 25 mg/l, adsorbent dose of 1 g/l and initial pH = 4. The results

showed that although the equilibrium time for maximum ions removal was the same in single and binary ions systems, but the removal percentages attained was higher in single than in binary ions systems. The maximum removal of Al(III) in its single ion solution was 82% and decreased to 54% in the binary aluminum-iron solution. Similarly, the maximum Fe(III) removal was 67% in the single iron solution and decreased to 62% in binary aluminum-iron solution system. Similar findings were obtained by Lugo-Lugo et al. [13] for the biosorption of chromium and iron from their single and binary systems. Also in agreement with these results Uslu and Tanyol [14] stated that the presence different components in the solution creates a competition for the adsorption sites on the surface and some sites are occupied by the second component. As a consequence, the first component has a smaller parking space and its uptake is decreased.

In order to analyze the adsorption kinetics of Al(III) and Fe(III) adsorption by RHAC in single and binary systems, pseudo-first order, pseudo-second order and Elovich kinetic models were applied to the experimental data.

The pseudo-first order model is represented by: [15] . The pseudo-second-order equation [16] based on equilibrium adsorption is expressed as:. Finally, according to El-Ashkouty [17] the simple Elovich’s model is expressed as. Where k₁ (min⁻¹) is the pseudo first order adsorption rate coefficient, k₂ (g/mg∙min) is the rate constant of second-order adsorption, q_e and q_t are the values of amount adsorbed per unit mass at equilibrium and at any time (t), respectively. The parameter α is the initial sorption rate constant (mg/g∙min) and β is related to the extent of surface coverage and activation energy for chemisorptions (g/mg).

The kinetic constants and the determination coefficients for the three tested models are summarized in Table 2. The results obtained were interpreted by considering the determination coefficient values obtained and also by comparing the values of the adsorption capacities calculated from the models (q_e, calc.) with those obtained experimentally (q_e, exp.).

The results indicated that in both single and binary systems, the adsorption of Al(III) and Fe(III) was found to be best described by the pseudo-second order model as well as the Elovich’s model. On the other hand, the pseudo first order model showed the least fit to the data. The adequacy of the pseudo second order model to fit the adsorption data suggests that the rate-limiting step is a chemical sorption or chemisorptions involving valence forces through sharing or exchange of electrons between adsorbate and adsorbent [18] . Moreover, the applicability of both pseudo-second-order and Elovich kinetic models to the experimental data gives a further confirms that chemisorptions is the dominant process in controlling [19] .

The results also showed that there was a noticeable reduction in RHAC adsorption capacity values towards both studied ions in the binary system as compared to the single systems which seems logic referring statistical vacant active adsorbent sites capacity of RHAC.

Based on previously published studies, the effect of pH on the removal of aluminum and iron ions by RHAC

was studied in the range between pH 2 and 5 to avoid metal ions hydroxide precipitation at higher pH values [20] and [21] . Figure 3 represents the removal percentage of aluminium and iron ions at different initial pH values.

It is clear that the removal of both metal ions was low at pH 2 and increased by increasing the pH value of the metal ions solution. It is generally agreed that at very low pH values the metal ions removal is weak due to the competition between the positively charged hydronium ions present in solution and the metal ions to occupy the adsorbent’s active sites. As the solution pH is raised less hydronium ions are present and thus the opportunity of the positively charged metal ions to occupy the adsorbent’s surface increases. The maximum removal of aluminum and iron ions due to adsorption was obtained at pH 5 and pH 3, respectively (Figure 3). Thus a middle pH value (pH = 4) was used in all subsequent experiments.

In agreement with our results; Cayllahua and Torem [22] stated that at pH 5 the optimum removal of aluminum ions onto Rhodococcus opacus was obtained. The pH 4.5 was selected as the optimum pH for aluminum ions removal by P. pavonica biomass [23] . Similarly, Kousalya et al. [24] reported pH 3 to be the optimum pH for Fe(III) by nano-biocomposites. Also Quintelas et al. [25] reported pH 3.5 as the optimum pH for ferric ions removal onto E. coli biofilm supported on kaolin.

The adsorption of Al(III) and Fe(III) by RHAC was analyzed by well documented Langmuir (Equation (4)) [26] , Freundlich (Equation (5)) [27] , Temkin (Equation (6)) [28] and Dubinin-Radushkevich (Equation (7)) [29] isotherm models.

where: C_eq: equilibrium concentration in mg/l; q_e: equilibrium capacity in mg/g; q_max: maximum capacity in mg/g; b (l/ mg) is Langmuir constant; k_f and n are Freundlich constants; b_t and a_t are Temkin isotherm constants; q_m is the theoretical saturation capacity (mg/g), β is a constant related to the mean free energy of adsorption per mole of the adsorbate (mg²/J²), and ε is the Polanyi potential.

The adsorption free energy (E; kJ/mol) can be calculated from the D-R model as follows: E = 1/√−2β.

All the parameters calculated from the studied equilibrium models along with the correlation coefficients (R²) are given in Table 3.

The coefficients of determination (R²) for the Langmuir plots (Figure 4) were found to be ≥0.98 for both aluminium and iron ions in their single and binary solutions. It can be concluded that the adsorption of the studied metal ions onto RHAC is well fitted to the Langmuir model. These results suggest that the adsorption of

Al(III) and Fe(III) onto RHAC is a monolayer adsorption taking place at the surface groups binding sites of the adsorbent. The maximum adsorption capacities calculated from the Langmuir model were found to be 34.48 and 45.45 mg/g for aluminium and iron in their single solutions and 21.28 and 26.32 mg/g for aluminium and iron in their binary solutions, respectively. It is clear that the maximum capacity of the metal ions in their single solutions was higher than the maximum capacity values recorded in the binary ions solution. A comparison between the maximum Langmuir capacity values obtained in the present study with those previously reported in literature for aluminium and iron adsorption onto various adsorbents is given in Table 4.

The Freudlich model plots obtained for the experimental data are given in Figure 5. The R² values of the Freundlich model (0.97 and 0.96 for both aluminium and iron ions in their single binary solutions, respectively) were found to be lower than those obtained for the Langmuir model. These results suggest that the adsorption of Al(III) and Fe(III) onto RHAC is not well described by the Freundlich model compared to the Langmuir model.

The plots of Dubinin-Radushkevich (D-R) isotherm are shown in Figure 6. The D-R model showed high determination coefficient values (R² ≥ 0.97) suggesting a good fit to the experimental data. The adsorption free energy (E; kJ/mol) can be calculated from the D-R model as follows: E = 1/√−2β. The E (kJ/mol) value gives information about the physical or chemical nature of the adsorption mechanism. If it lies between 8 and 16 kJ/mol, the adsorption process takes place chemically and while E < 8 kJ/mol, the adsorption process proceeds physically [30] . As seen from Table 3, the values of E for both aluminium and iron ions were found to be in the range 8 - 13 kJ/mol. Thus, the calculated values of E suggest that the adsorption process is chemisorptions.

On the other hands the Temkin isotherm model (Figure 7) did not show any fitness to the experimental data of Al(III) and Fe(III) adsorption onto RHAC with R² values < 0.80.

Based on the R² values which is a measure of the goodness of model’s fit [37] , it can be concluded that the adsorption data of aluminium and iron ions fitted well to the Langmuir model followed by the D-R model.

The results also show that the maximum capacity of the metal ions in their single solutions was higher than the maximum capacity values recorded in the binary ions solution.

The activated carbon prepared from rice hull was characterized by FTIR and SEM techniques in order to identify the surface functional groups and surface morphology of the adsorbent.

The FT-IR spectrum of (RHAC) is shown in Figure 8. It can be observed that the spectrum displays a number of bands corresponding to several functional groups on the activated carbon surface that may facilitate the adsorption process. The spectrum of RHAC showed that a broad band at 3423 cm⁻¹ this band is characteristic of the OH stretching vibrations mode of hydroxyl functional groups including hydrogen bonding of chemisorbed water or may be due to binding of OH group with polymeric structure of rice hulls [38] . The band at 2922 cm⁻¹ could be attributed to C-H stretching of aliphatic carbon or it can be due to CH₂ or CH₃ deformation [38] . The peak at 1625 cm⁻¹ related to C=C stretching for unsaturated aliphatic structures [39] . The bands at 1099 and 798 could be due to Si-O-Si stretching and Si-H groups, respectively [40] .

The scanning electron microscopy micrograph of RHAC is shown in Figure 9. The figure shows a variety of cavities as well as a porous morphology with pores of different shapes and sizes. This irregular surface format leads to increasing the contact area, which facilitates the pore diffusion during adsorption [41] .

Thus by considering the FTIR surface bands and the surface morphological characteristics of RHAC, it can be inferred that RHAC possess a surface capable of adsorbing metal ions.

The results of real water samples treatment using RHAC and CAC are presented in Table 5. It can be observed that in all cases (RHAC) showed higher removal efficiency for aluminium and iron ions as compared to the

commercial activated carbon. It is worth to note that from an economic point of view, the use of RHAC will be more advantageous than the commercial activated carbon as RHAC is cost effective natural adsorbent.