Heavy metals are the major cause of water pollution. Among many hazardous toxic heavy metals which cause significant health risks for human, lead stands on the second   . Hence, there is a need for the development of simple and cost effective methods to remove it. The most familiar methods for the effective removal of heavy metals from waste waters include precipitation, ion exchange, membrane filtration, electrochemical technologies and adsorption. Among many such methods, adsorption is the best choice as it offers easy operation and the regeneration of the adsorbents with desorption processes for multiple use.
Heavy metals remediation has been efficiently accomplished in the past by several nano metal oxides, including: ferric oxides, manganese oxides, aluminum oxides, titanium oxides, magnesium oxides and cerium oxides  . This could be attributed to large surface areas and high activities caused by the size quantization effect in nanoparticles  . Of those ferric oxides nanomaterials hae gained increased attentions because of the ease of regeneration of iron oxides using magnetic field; after regeneration it is possible to use it reputedly. Such easy regeneration is important to improve the operation efficiency and decrease the cost during treatment of water/wastewater  . In addition to this nano-sized ferric oxides have been proved to be the first adsorbents for toxic metal sorption which is possibly due to their large surface area   . And it can be used directly to contaminated areas with insignificant dangers of secondary contamination  . The past studies revealed the existence of various forms of ferric oxides in nano-sized iron oxide for heavy metals removal from water/wastewater includes, amorphous hydrous iron oxides, hematite (α-Fe2O3), goethite (α-FeOOH), maghemite (γ-Fe2O3) and magnetite (Fe3O4)   .
The synthesis routes of metal oxides nano materials include gas phase deposition, electron beam lithography, pulsed laser ablation, laser induced pyrolysis, powder ball milling, aerosol from physical methods and reverse micelle (or micro-emulsion), sol-gel, co-precipitation, hydrothermal, electrochemical deposition, sonochemical, thermal decomposition, gas-phase reduction from chemical methods  . Among all the synthetic methods, co-precipitation  , thermal decomposition and/or reduction  and hydrothermal synthesis  methods were widely used and are easily scalable with high yields. The main draw backs of previous studies were the use of toxic organic solvents. In this context, ongoing research has modified the existing different techniques by using only water as solvent.
For this work co-precipitation has been used for synthesis of the adsorbent by using only distilled water as a solvent. The main objective of this study was to synthesize and characterize the as-synthesized materials using different analytical techniques (FT-IR, XRD, SEM-EDX and TGA-DTA) and to evaluate its adsorption efficiency for the removal of lead ions from aqueous solution. For this study the mechanism of the adsorption was evaluated with the help of Langmuir, Freundlich, D-RK and FH adsorption isotherm models. Moreover to check the kinetics of adsorption PFO, PSO and Elovich models were used for adsorption-reaction and IPD model were used for adsorption-diffusion.
2.1. Synthesis of Nano Sized Iron Oxide
All the chemical/reagent used for the current investigation was analytical grade and only distilled water was used as a solvent. Nano-sized Fe oxide was prepared by co-precipitation method using iron (III) nitrate Nonahydrate [Fe (NO3)3∙9H2O] salt as precursor and cetyltrimethyl ammonium bromide (CTAB) as a precipitating agent. The respective ratio used during mixing of precursor and precipitating agent was 8:1 with continuous stirring on magnetic stirrer. Sodium borohydride was then added as reducing agent. The pH of the solution was adjusted to 12 by drop wise addition of acid and base. The solution was refluxed at 100˚C for 2 hours with constant stirring. The obtained precipitate was then separated from solution using centrifuge. After washing the precipitate was washed with alcohol and water, it was dried in oven over night at a temperature of 105˚C and finally the dry powder was calcined at 450˚C for two hour.
2.2. Batch Adsorption Experiments
The sorption test was conducted in 50 mL erlenmeyer flasks containing adsorbent to solution ratio of 0.5:300 [0.05 g of iron oxide: 30 mL of 45 mg/L of Pb(NO3)3]. The pH of the solution was adjusted with the help of dilute HCl and NaOH solution. Equilibration of the experiments was done on a rotary shaker. The amount of lead adsorbed was known by the differences in equilibrium and initial lead concentration. Atomic adsorption spectrometer (AAS) was used for quantitative estimation of lead. The samples containing more lead ion than the highest concentration of detection limit of the instrument were diluted and the dilution factor was compensated during final calculation. The pH, adsorbent dose, agitation speed, contact time and initial lead ion concentration were optimized during experimentation  .
From mass balance for the adsorbate in the glass ware is:
From which a relationship between value of C and the corresponding equilibrium value of q can be established. To determine equilibrium relationship qo become equal to zero (qo = 0),
The percent of adsorption (%) were calculated using equation:
where: Co = initial concentrations (mg/L) and Ce = equilibrium concentrations (mg/L) of lead ion, qe = adsorption capacity of adsorbent (mg/g), V = volume of reaction mixture (L), m = mass of adsorbent (g).
2.2.1. pH Optimization
To optimize the effect of pH on the lead sorption, 0.1 g of the sorbent was added into 50 mL erlenmeyer flask containing 30 mL of 45 mg/L lead ions, then by varying the pH of the solutions from 1 - 11 with two increments and keeping the other conditions at constant value (rotation speed of 120 rpm and contact time 130 min.) the optima were obtained.
2.2.2. Dose Optimization
Adsorbent dosage was evaluated and optimized by taking various amounts of adsorbent dose as, 0.01, 0.03, 0.05, 0.1, 0.2 and 0.4 g, while the other parameters were kept at constant value and pH was maintained at the optimized value of 9.
2.2.3. Agitation Speed Optimization
The effect of agitation speed was optimized by varying the speed of rotary shaker to 30, 90, 120, 150, 180 and 200 rpm. During here during optimization dosage and pH were kept at optimized value and initial lead concentration at constant value.
2.3. Adsorption Isotherms and Kinetics of Adsorption
The degree of affinity of the adsorbate towards adsorbent is quantified using adsorption isotherms. A number of isotherms have been developed to describe equilibrium relationships. Here Langmuir, Freundlich, D-RK and FH models were engaged to describe the experimental results of lead sorption. Sorption isotherms experiment was done again by keeping all parameters at optimized values and by varying lead ion concentration as: 25, 45, 65, 85, 105 and 125 mg/L in separate 50 mL erlenmeyer flask. And the study of kinetics was conducted by taking various contact time 10, 50, 90, 130, 170, 210 and 240 minutes under optimized values of pH, adsorbent dose, agitation speed and lead ion concentration.
2.4. Thermodynamics Study
In order to determine the effect of temperature on sorption phenomenon, all predetermined and optimized values of parameters were used and the temperature established during sorption was varied from 25˚C to 55˚C with increment of 10˚C.
2.5. Reproducibility Studies
Lead ions desorption studies were done using lead ions loaded powder sample which is obtained after adsorption of lead ions on powder using all optimized values. 0.05 g of iron oxides of lead loaded powder was added into flasks containing 30 mL of double deionized water. 0.1 M NaOH and 0.1 M HCl solutions were used to adjust pH of the solution to: 1, 3, 5, 7, 9 and 11. After agitation the filtered solution were analyzed for desorbed lead ions concentration.
3. Result and Discussion
The as-synthesized iron oxide nanoparticle which is examined by XRD technique was given in Figure 1(a). The obtained approximate crystalline size of the nano particle was 16.55 nm. to calculate the size of the particle Debye Scherer equation were used, ; were D is the mean size, K is constant (0.94), λ is wavelength of X-ray (0.15506 nm), β is excess line of broadening and θ is Bragg angle, β = G - g; were G is the line width (in radians) and g is instrument line broadening (in radians)  . All peaks on diffraction pattern spectrum confirm the structure to be rhombohedral α-Fe2O3 in comparison other results  . Width of the peaks confirms the low particle size or high surface area to low volume ratio values, but, the low strength of peaks reveals that the as-synthesized nano-materials may not be fully crystalline. The 2θ values and the respective planes were found to be, 24.42 (012), 33.44 (104), 35.78 (110), 41.28 (113), 49.82 (024), 54.39 (116), 57.55 (018), 62.82 (214) and 64.46 (300)   .
SEM (Model ZEISS EVO 18 with INCA software for quantitative Analysis) is another useful tool for the morphological analysis of the surface of solids. The images of as-synthesized nano material with different magnification are as shown in Figure 1(b). These images confirm the presence of round shaped iron oxide nanoparticles. EDX proves that the presence of iron (Fe), oxygen (O) and carbon (C). The appearance of carbon on EDX was from standard used during analysis and the weight percent and atomic percent of the element are given on the spectrum. Except, Fe and O, no other elements were found in the spectrum.
The FTIR (Bruker IFS120 M Perkin Elmer) spectrum (Figure 1(c)) reveals strong bonding interaction between Fe and O in iron oxides nanoparticles. The broad vibrational peaks observed at 3400 and 1620 cm−1 corresponds to normal vibrations of water (H-O-H) and hydroxyl groups (O-H) absorbed on the surface, respectively  . The peaks at 478 and 544 cm−1 are due to bending vibration of Fe-O and Fe-O-Fe  . The other appeared peaks may be due to solvents used during synthesis.
The result of stability test of the as-synthesized iron oxide nanomaterial which is done by TG-DTA (DTG-60H) instrument is shown in Figure 1(d). Thermogravimetric analysis indicates two different weight losses observed at around 258 and 660˚C. It was observed that only 0.255 mg of nanosorbent was lost out of 11.749 mg, therefore, the stability of iron oxides is believed to be good.
The effect of pH, adsorbent dose, speed of agitation and contact time on the lead ion sorption efficiency of nanosorbent is presented in the Figure 2.
Figure 1. (a) XRD (b) SEM images at different magnification and EDX spectra (c) FT-IR spectrum (d) TG-DTA Thermogram of iron oxide calcined at 450˚C.
On pH optimization it was observed that greatest lead ion adsorption occurred under basic condition and the optimum pH value for sorption was found to be 9 as revealed by pH verses % of sorption plot (Figure 2(a)). The lead removal from the solution increased with pH, because increasing solution pH favors the de-protonation of nano sorbent surface, and this leads to increasing the negatively charged sites and hence the electrostatic forces of attraction between lead ions and the adsorbent increases. On lowering the pH, it has been reported that competitive adsorption occurs between the competing metal ions and H+ ions in the solution  . The optimum dosage value was determined to be 0.2 g (Figure 2(b)) and beyond this value, the rises in amount of adsorbent dosage give no significant role on adsorption due to less availabilities of adsorbate and also it may be due to the formation of agglomeration between the particles. The agitation speed optimization that assists in diffusion of lead ion to adsorbent was shown in Figure 2(c); the obtained optimum point was at 90 rpm. After the optima, the decreasing behavior of the plot shows desorption of lead ions from physically adsorbed surface due to increase of agitation speed. The optimum value of contact time obtained was 170 minutes (Figure 2(d)); after the optimum
Figure 2. (a) Effect of pH (b) adsorbent dose (c) speed of agitation (d) contact time (e) initial lead ion concentration on adsorption of lead ion by iron oxide.
point since the adsorbent sites almost occupied it becomes constant. Short adsorption time for iron oxide based nano-adsorbents may possibly due to porosity of the adsorbent.
3.6. Adsorption Isotherms
The effect of initial concentration on lead ion uptake shows almost constant flow up to optimum value (45 mg/L) (Figure 2(e)) for two hours  . This trend is possibly due to the availability of active sites which gradually decreases as those sorption sites were used up and it prevents more ion adsorption by adsorbent specially if Langmuir isotherm model is the controlling mechanism  . The lead ion adsorption efficiency by the nanosorbent in terms of percentage at optimum point was found to be 98.86%.
The linear Langmuir equation is as given below:
The significant characteristics of a Langmuir isotherm can be expressed in terms of a constant separation factor or equilibrium parameter, RL:
The linear Freundlich equation expressed as:
Log qe = log Kf + 1/n log Ce (6)
The linear Dubinin-Radushkevich equation:
, where and (7)
The linear Flory-Huggins equation:
where, Co = initial adsorbate concentration in solution (mg/L), Ce = adsorbate equilibrium concentration in solution, qo = initial amount of adsorbate per unit mass of adsorbent (mg/g), qe =amount of adsorbate accumulated per gram of the adsorbent material, qmax = maximum uptake corresponding to the site saturation, b = ratio of adsorption and desorption rates, Kf = distribution coefficient and represents the quantity of adsorbate adsorbed onto adsorbent for unit equilibrium concentration, 1/n = an empirical constant related to the magnitude of the adsorption or surface heterogeneity, is Fractional coverage, ε = Dubinin - Radushkevich isotherm constant and qs is saturation capacity (mg/g), β is the constant related to free energy, n is the number of ions occupying adsorption sites, T absolute temperature (k) and R is the universal gas constant (8.314 J/molK). From sorption isothermal studies, the obtained sorption parameters given in Table 1 were calculated from the slopes and intercept of the linearized Langmuir (Figure 3(a)) and Freundlich plot (Figure 3(b)), D-RR (Figure 3(c)) and FH (Figure 3(d)) adsorption isotherm model. The correlation coefficient of Langmuir sorption isotherm model is much higher than the other models, which show that, Langmuir sorption isotherm model fits the sorption data better. The RL value indicates the type of the isotherm to be unfavorable if RL > 1, linear if RL = 1, irreversible if RL = 0 or favorable if RL < 1  . The RL values for all concentration were found to lie between 0 and 1 indicates favorable sorption. The value of 1/n being in between 0 and 1 on Freundlich model indicates the heterogeneity of the surface, to be more heterogeneous its value should be closer to zero. The value of 1/n for this study is 0.352 not as such closer to zero, so the sorption mechanism deviates more towards Langmuir adsorption. The magnitude of n being in between 2 - 10 confirms the favorability of sorption process and if its value is below 1 becomes chemical adsorption and if it is greater than one become physical  . Being values of n for this study 2.86 states the reaction was under control of physical adsorption.
From D-RK model the calculated mean adsorption energy, E was 1.89 KJ/mol, being its value less than 8 KJ/mol specifies the domination of physical interaction.
Table 1. Adsorption isotherm constants of different models for lead ion adsorption.
Figure 3. (a) Langmuir; (b) Freundlich; (c) Dubinin-Radushkevich; (d) Flory-Huggins adsorption isotherm of lead ion on Fe oxidesadsorbent.
And from FH model the obtained values for free energy, ΔG was −21.36 (kJ/mol), this shows that the spontaneity of the reaction   . The obtained maximum sorption capacity (qmax) for the sorbent is 70.422 mg/g.
3.7. Kinetics of Sorption
The result of kinetics of sorption study obtained after optimization of all parameters is shown in Figures 4(a)-(d). And the results of kinetics study and kinetics constants values are given in Table 2. From different kinetic models used to interpret the time dependent experimental data, the linear equation of the PFO, PSO, Elovich and IPD equations were used and the respective equation become as follows:
Table 2. Kinetic parameters and correlation coefficient values of different kinetic models.
*calculated value and **experimental value.
Figure 4. (a) PFO (b) PSO (c) Elovich (d) IPD kinetics model on removal of Pb(II) by Fe oxide.
Weber-Morris IPD model:
where, qe and qt are the amounts of adsorbate adsorbed on the adsorbent at equilibrium and at various times t (mg/g), respectively, k1 is the rate constant of the FSO model for the adsorption process (min−1), k2 is the rate constant for the PSO model (mg/gmin, α is the initial sorption rate (mol/g∙min) and β is the desorption constant (g/mol)), ki is the IPD rate constant (mg/g∙min1/2).
Depending on R2 values relatively, PSO models adequately described the kinetics of sorption of lead ion better than others model, and its theoretical equilibrium capacity of 26.68 mg/g fit well with experimental data value of 27.17 mg/g, but not for PFO. Elovich model which is useful model in describing the chemisorption behavior of adsorbate-adsorbent interaction gives comparatively less R2 values (Figure 4(c)). Regression coefficient value of intra-particle diffusion plots clearly demonstrates its deviation from linearity. To say the reaction is under the control of IPD the line should pass through the origin, unlike that of obtained result on this paper (Figure 4(d)). Therefore, possible to deduce that the reaction is totally under control of surface/adsorption-reaction mechanism, this indicates that the rate determining step become adsorption-reaction interaction instead of adsorption-diffusion   .
The effects of temperature on lead ion sorption are shown in Figure 5(a). The thermodynamic parameters (∆G, ∆H and ∆S) can be calculated by using the following equation:
where R (8.314 J/mol∙K) is the gas constant, T (K) is the absolute temperature and Kc is the standard thermodynamic equilibrium constant which is defined by qe/Ce.
By plotting the graph of lnKc versus T−1, the value of ∆H˚ and ∆S˚ can be estimated from the slopes and intercept. The obtained values of ΔG˚, ∆H˚ and ∆S˚ different temperature were given in Table 3. The sorption rate for Pb(II) ions was found to increase with increase in the temperature due to increase in the surface area of adsorbent and diffusion of it. Consequently, the numbers of sites were found to increase because of the breaking of domestic bonds which in turn increases lead ions adsorption from aqueous solution. The decrease of ∆G values with the increasing temperature shows that the sorption was favorable at higher temperature. The positive values of ΔH indicates the endothermic nature of Pb(II) sorption process. The positive values of ∆S show that lead ions in liquid phase (aqueous solution) exhibits more disordered distribution than that in relatively ordered state of solid phase and during adsorption the solid-liquid interface become more random  .
3.9. Reproducibility Studies
The reusability of the adsorbent after its repeated usage has been a crucial factor for economic compatibility. From the reproducibility/desorption studies (Figure 5(b)),
Table 3. Thermodynamic parameters for lead ion sorption.
Figure 5. (a)Thermodynamics; (b) Effect of pH on desorption of lead.
the amount of desorption of lead ion decreases as the pH of solution increases. 83.94% of Pb(II) ions was removed in the first cycle. Desirability was obtained using equation:
% Desorption Efficiency = D/A × 100 (16)
where, D = Desorbed is the concentration of lead ion after the desorption process and A = Adsorbed is (Co - Ce) for each recovery process.
Co-precipitation method has been successfully carried out to synthesize nano sized iron oxide at a relatively low temperature. XRD pattern confirmed the rhombohedral (hexagonal) structure of α-Fe2O3. The crystalline size of iron oxides particles was found to be 16.55 nm. The as-synthesized iron oxide nanoparticles was examined for morphological details by SEM which confirmed almost round shape and nano-powders were found to be less agglomerate. The observed peaks in EDX spectrum confirmed the presence of Fe & O and the absence of impurities in the prepared Fe2O3. The presence of Fe-O bond and its stretching vibration mode was confirmed by FTIR data. Langmuir isotherm model describes the adsorption data well compared to Freundlich, D-RK and FH models. From PFO, PSO, Elovich adsorption-reaction models, PSO fits well and IPD adsorption-diffusion models fail to fit and the line also will not pass through the origins. This means that the reaction is under control of adsorption-reaction model. The obtained maximum sorption capacity for the sorbent was 70.422 mg/g and 83.94% of Pb(II) was liberated during desorption studies.
Authors are great full to the management of Adama Science and Technology University for providing financial support towards this research work. The author acknowledges Mr. Yilkal Dessie for software support and Dr. Senthilkumar Subramanian for characterization support.