The contamination of groundwater with heavy metals particularly iron has been reported  . Iron is slowly released into ground water as naturally occurring and weathering iron bearing minerals and rocks infiltrating through the underlying formations are dissolved in the water and accumulates in the aquifers that serve as sources of ground water. Various products of industrial and human activities: Industrial effluents, acid-mine drainage, sewage, landfill leachate and agricultural activities also contribute iron to groundwater  . High content of iron in groundwater has been reported  . Although iron is an essential mineral for human, concentrations in ground water above the stipulated safe allowable limit makes ground water turbid, affects its domestic use (taste, color, odor) and affects its industrial applications. Oxidation of dissolved iron in water converts it to a red-brown solid which stains laundry and plumping fixtures, glassware, enhances the growth of iron bacteria, which forms dark-colored slime layers on the inner side of pipes and enhance corrosion of water pipes; excess iron deposits in the heart muscle can cause heart failure as well as abnormal heart rhythms  .
Groundwater is the major source of industrial, domestic and drinking water supply in Nigeria  and major cities of Rivers State. UNESCO  , reported that, more than five million people died every year in developing countries due the diseases associated to Water borne diseases. This has necessitated the stipulation of safe allowable maximum concentration limit for iron (0.3 mg/L) in ground water for drinking purpose by the World Health Organization (WHO)  . Treatment of the ground water supply is therefore recommended particularly when concentrations are above the stipulated allowable safe maximum.
Numerous processes abound for the removal of iron and indeed heavy metals from water and waste discharges. These methods include: solvent extraction  , chemical precipitation  , membrane process  , adsorption  and ion-exchange  . However, Adsorption processes have been reported to be an efficient economical method and the primary methods for heavy metal removal  , the other methods are more expensive and sometimes inefficient especially when the toxic heavy metals are present in low concentrations. Adsorption technique has found extensive applications in deodorization, purification of drinking water, treatment of waste water, separation of various organic and inorganic chemicals and the adsorption of heavy metals from aqueous solutions  using various materials as adsorbents: zeolites  , silica  , resins  , biological adsorbents  and activated carbon   .
Adsorption using activated carbon as adsorbent has been found to be superior to other techniques because of its simplicity of design and capability to adsorb a broad range of pollutants or contaminant efficiently from aqueous or gaseous media  . Activated carbon (AC) is a porous material with exceptionally high surface area, large pore volume and well-developed internal micro porosity and wide spectrum of surface functional groups  . The use of low cost precursors such as agricultural wastes that are available in large amounts, renewable and with high carbon content for the sorption of heavy metals from solutions have been reported: rice hulls  ; cotton stalks  ; wood sawdust  ; palm kernel shell  ; almond shells, olive stones, and peach stones  ; apple pulp  ; sugarcane bagasse  and periwinkle shells  . Periwinkle shells have been reported  to have good adsorption capacity because of its polar functional groups (primary amines, hydroxyl, carboxylic acid, amide and phenolic groups. Periwinkle shell is an abundant waste material found in the riverine parts of the Niger Delta where its content (Lymneanetalensis) is an edible seafood.
The objective of this research is to investigate the use of periwinkle shells as an adsorbent for the removal of iron ions from groundwater. The activated carbon was produced from periwinkle shells (carbonization and activation with tri-oxo nitrate (V) acid), characterized and adsorption experiments of iron (III) from its aqueous solution performed. This was used to study the kinetics, equilibrium and thermodynamics of the adsorption process, thus generate the required data for parameters used in designing adsorber units for water purification.
2. Materials and Methods
The Periwinkle shell was obtained from Mile 3 market in Port Harcourt, Iron (III) oxide (Fe2O3) and the activation agent, nitric acid (HNO3) were purchased from a chemical shop at the industrial chemical section of Mile 3 market, while distilled water was obtained from the Department of Chemistry laboratory, Rivers State University. All chemical reagents used were of analytical grades. The pyrolysis reaction was performed using the “pyrolysis setup” in the reaction kinetics laboratory of the Department of Chemical/Petrochemical Engineering, Rivers State University, Port-Harcourt, Rivers State, Nigeria. The activation of the carbon, characterization of the produced activated carbon and the batch adsorption experiments were also performed at the reaction kinetics laboratory.
The periwinkle shell samples were washed several times with distilled water to remove sand and other surface impurities and sun dried.
The periwinkle shells were heated in a muffle furnace for 30 minutes to dry (free of water). A measured weight of the shells was introduced into the reactor and pyrolized at 300˚C for three hours in the absence of air. A receiver was connected to the condenser to receive the distillates formed during the pyrolysis. The char material was cooled at room temperature before discharging into containers.
The carbonated material (carbon) from the periwinkle shells was crushed into powdered form using a crusher. A measured weight of the crushed sample was soaked in nitric acid of known concentration in a conical flask and stirred until the mixture turned a paste. The paste was heated in a muffle furnace at 800˚C for 2 hours in the absence of air to increase the surface area of the charred sample; the activated carbon was cooled at room temperature, washed with distilled water until pH was approximately 7 (no change in color when tested with red litmus paper) indicating no trace of the nitric acid used. The char was dried for 6 hours, sieved with sieves of various sizes (75 um, 150 um, 212 um, 300 um, 600 um to obtain activated carbon of various particle sizes) and were stored in tight nylons.
Characterization of Produced Activated Carbon
The produced activated carbon was characterized by determining selected physiochemical properties such as oil yield, pH, carbon yield, moisture content, ash content, pore volume/porosity and bulk density following standard procedures as outlined in  . All results were the average of duplicate analysis.
Preparation of Heavy Metal (Iron (III) Oxide Fe2O3) Solution
The stock solution of Fe(III) was prepared by dissolving required quantity (80.7 g) of (Fe2O3) in known volume of distilled water. Different initial concentrations of Fe(III) solution was prepared by appropriate proper dilution of the stock solution.
Batch experiments were performed by adding known amounts of produced activated carbon to 50 mL of the iron III stock solution in a 100 mL conical flask. The conical flask containing the adsorbent (PSC) and the stock solution was placed in a rotary shaker and shaken at 150 rpm at room temperature (32˚C) at different time periods at intervals of 20 minutes until equilibrium was attained(constant concentration of Iron (III) ion in the filtrate). The activated carbon was separated from the stock solution at each time interval by filtration using a Whatman number 41 filter paper. The concentration of iron (III) ion in the filtrate was analyzed at each time interval using an atomic absorption spectrophotometer (AAS). The Batch adsorption experiments were replicated trice.
The percentage of metal ion removed at each time interval was obtained using the relationship:
The amount of metal ion adsorbed by the adsorbent at each time (t) was evaluated using the equation below;
The adsorption capacity of the adsorbent was also calculated using the formula:
where: Co = Initial concentration of Iron III ion present in waste water before adsorption.
Ct = Final concentration of Iron III ion present after adsorption at time t.
Ce = Concentration of Iron III ion in the filtrate when equilibrium was attained.
V = Volume of stock solution (mL).
W = Mass (g) of adsorbent used.
Effect of Process Parameters on Adsorption Rate
The effect of contact time on adsorption was studied by conducting the batch adsorption experiment with a known particle size and dose (weight) of adsorbent at different contact times from 20 to 120 minutes at 20-minutes interval. The effect of particle size on adsorption was studied by conducting the batch adsorption experiment with a known weight of adsorbent and fixed contact time for different particle sizes of 75 µm, 150 µm, 212 µm, 300 µm and 400 µm of the activated carbon. The effect of adsorbent dose on adsorption was studied by conducting the batch adsorption experiment with a known particle size and contact time for different dosages of adsorbent (0.2 - 0.8 g). The effect of pH of stock solution on adsorption was studied by conducting the batch adsorption experiment with a known particle size, dose (weight) of adsorbent, fixed contact time at different pH values (1 - 5) of the stock solution. The pH of the solution was varied by dilution of the stock solution or by adding known volume of 0.1 M HCl or 0.1 M NaOH to stock solution as appropriate.
Kinetic analysis was performed to investigate the rate of adsorption, determine the mechanism of adsorption and the potential rate controlling steps (mass transport, pore diffusion or chemical reaction). Kinetic models have been used to correlate experimental data  . The pseudo first and second order kinetic equations were fitted to the experimental data to model the adsorption kinetics of Iron (III) adsorption onto periwinkle shell activated carbon. Kinetic models were used for goodness of fit of the experimental data using the correlation coefficient (R2) as a measure of agreement between the experimental data.
Pseudo-First Order Kinetic Model
This model is associated with physical adsorption where the adsorption process is controlled by weak interactions between the adsorbate and the adsorbent surface  .
The pseudo-first order rate equation by the Lagergren is given as:
where and are the adsorption capacities at equilibrium and at time t (mg∙g−1) respectively and is the pseudo-first order adsorption rate constant (L∙min−1). Integration at: to and , gives the linear form of the equation as:
The values of and were obtained from the slope and intercept of the plot of versus t respectively.
The Pseudo-Second Order Kinetic Model
This model is based on the assumption that the adsorption is controlled by chemical adsorption-chemisorption (the rate of direct adsorption/desorption process seen as a kind of chemical reaction)  . The pseudo-second order kinetic model is given  as:
Integration at: to and , the linear form of the equation was obtained as:
where is the pseudo second order rate constant. The values of model parameters, were calculated from the slope and intercepts of
the linear plot of vs t.
Adsorption Equilibrium Studies
Adsorption isotherms shows the relationship between the amount of adsorbed metal ions per unit of bio sorbent (adsorbent) ( ) to the metal concentration in solution ( ) at equilibrium at a given temperature, pressure, pH and total solute concentration  . Adsorption isotherm study was carried out at four different temperatures: 30˚C, 40˚C, 50˚C and 60˚C. The sorption equilibrium data from the experiments were fitted with the Langmuir, Freundlich, and Temkin isotherms models to describe the adsorption isotherm of Iron (III) ion on periwinkle shell activated carbon.
Langmuir Adsorption Isotherm
The Langmuir isothermmodel is based upon an assumption of monolayer adsorption (at maximum adsorption only a saturated monolayer of solute molecule is formed on the adsorbent surface) onto a homogeneous surface containing finite number of adsorption sites of uniform energies of adsorption, all adsorption occurs through the same mechanism with no transmigration of adsorbate in the plane of the surface (molecules of adsorbate do not deposit on other already adsorbed molecules of adsorbate, only on the free surface of the adsorbent)   . The Non Linear form of the Langmuir isotherm is given as:
The linearized form of this model is:
The constants and b are the Langmuir constants representing the maximum adsorption capacity for the solid phase loading (mg∙g−1) and the nature of adsorption and the shape of isotherm (L∙mg−1). These were evaluated from the
slope and the intercept of the plot of vs of the linear model. An
essential feature of the Langmuir isotherm can be expressed in terms of a dimensionless separation factor or equilibrium parameter ( ) defined  as:
where (mg∙L−1) is the initial amount of adsorbate and b (L∙mg−1) is the Langmuir constant described above. The value of indicates whether the adsorption isotherm is unfavorable ( ), linear ( ), favorable ( ), or irreversible ( ).
Freundlich Adsorption Isotherm
The Freundlich isotherm model is based upon an assumption of multilayer adsorption onto a heterogeneous surface where the sorption energy distribution decreases exponentially  . The Non Linear form of this model takes the form:
The linearized form of the model is:
where: and 1/n are empirical constants related to adsorption capacity and adsorption efficiency. is the Freundlich heterogeneous adsorbent constant and is the amount adsorbed at unit concentration, (mg∙g−1), n is Freundlich constant and is a measure of the intensity of adsorption. Where n > 1 indicates that affinity decrease with increasing adsorption density. These constants were evaluated from the slope and intercept of the linear model plot of versus .
Temkim Adsorption Isotherm
The Temkim isotherm takes into account the interactions between adsorbents and metal ions to be adsorbed and is based on the assumption that the free energy of sorption is a function of the surface coverage where the heat of adsorption of all the molecules in a layer decreases linearly due to adsorbent-adsorbate interactions and that adsorption is characterized by a uniform distribution of binding energies, up to some maximum binding energy  .
The isotherm is as follows:
A plot of versus enables the determination of the adsorption isotherm constants B and from the slope and intercept respectively.
The thermodynamic behavior of the adsorption process was studied by determining parameters such as the Gibbs free energy change (kJ/mol×K), the standard enthalpy change (kJ/mol), and the standard entropy change (J/mol×K) using the Van’t Hoff equations:
The distribution adsorption coefficient/equilibrium constant (Ke) of a metal ion adsorbed by an adsorbent (mL×g−1) can be obtained from the equation given by  as:
where R is the universal gas constant (8.314 J×K−1), T is the absolute temperature in Kelvin.
Equilibrium concentration of adsorbed species in bulk phase (Ce) and adsorbed phase (qe) at different temperatures were appropriately correlated to estimate the thermodynamic parameters.
Adsorption experiments were conducted at different temperatures of 30˚C, 40˚C, 50˚C and 60˚C to obtain batch equilibrium data used in calculating the distribution adsorption coefficient/equilibrium constant ( ) from equation (16) at these temperatures. The plot of versus (Equation (14)) was used to calculate the values of and from the slope and intercept respectively. was then obtained from Equation (15) for all temperatures.
3. Discussion of Results
3.1. Characterization of Activated Carbon
The characterization results of the produced activated carbon are given in Table 1.
Table 1 shows the properties of periwinkle shell activated carbon in comparison with those of activated carbon produced from bamboo, coconut shell and palm kernel shell as reported by  . This showed that periwinkle shell activated carbon had more carbon yield, less water content which increases sorption rate  .
Table 1. Characterization of periwinkle shell activated carbon.
3.2. Effects of Process Parameters
3.2.1. Effect of Contact Time
The effects of contact time on adsorption (the percentage removal of Fe3+) using various particle sizes are shown in Figure 1.
Figure 1 shows that for all particle sizes considered, the percentage of Fe (III) ion adsorbed increased with contact time.
3.2.2. Effect of Particle Size
The effects of particle size on adsorption (the percentage removal of Fe3+) conducted at various contact times are shown in Figure 2.
Figure 2 shows that at all contact times of adsorption, the percentage of iron (III) ion adsorbed decreased with increase in particle size. The smaller the particle size of the activated carbon, the better the access to the surface area and the smaller the pore space which makes the activated carbon tends to retain more iron (III) ions.
Figure 1. Effect of contact time.
Figure 2. Effect of particle size.
3.2.3. Effect of Carbon Dosage
The effect of adsorbent dosage on adsorption (the percentage removal of Fe3+) is shown in Figure 3.
Figure 3 shows that as the dosage (quantity of activated carbon used) increased, the amount of iron (III) ion in the stock solution adsorbed (percentage adsorbed) by the activated carbon increased; an increase in adsorbent dose increases the surface area and number of active sites of the activated carbon  available to the solute (iron (iii) oxide) for adsorption, thus increasing the rate of adsorption  .
3.2.4. Effect of pH
The effect of pH of the stock solution on adsorption (the percentage removal of Fe3+) is shown in Figure 4.
Extraction of metal ions from aqueous media by adsorption had been reported  to be usually pH dependent because pH affects the surface charge of adsorbents and the degree of ionization. The result shows that as the pH of the stock solution increases the percentage metal iron removed increased. The increase in percentage removal can be explained by the fact that at higher pH, the adsorbent surface is deprotonated and negatively charged; hence attraction between the positive metal cations occurred  , thus increasing the adsorption rate.
Figure 3. Effect of carbon dosage.
Figure 4. Effect of pH.
3.3. Adsorption Kinetics
The parameters of the pseudo-first-order and pseudo-second-order models determined from the slopes and intercepts of Figure 5 and Figure 6 along with the corresponding correlation coefficients are listed in Table 2. The equilibrium amount adsorbed ( ) from the experiment ( ) and the calculated equilibrium amount ( ) from the two models are also shown in Table 2.
Figure 5. Pseudo-first-order-kinetics of Fe3+ ion adsorption onto PSC.
Figure 6. Pseudo-second-order-kinetics of Fe3+ ion adsorption onto PSC.
Table 2. Adsorption Kinetic parameters for the adsorption of Fe3+ of periwinkle shell carbon.
pseudo first order model with correlation coefficient R2 of 0.9338. The calculated (theoretical) value of from the pseudo second order model also agreed very well with the experimental values. Therefore, these results show that the pseudo second-order model was more reliable and accurate in correlating the experimental data compared to the pseudo first-order model. Hence the adsorption of Fe2O3 onto periwinkle shell activated carbon can be represented by a pseudo-second-order kinetic model which relies on the assumption that chemisorption may be the rate limiting step. In chemisorption, the metal ions stick to the adsorbent surface by forming a chemical (usually covalent) bond and finds sites that maximize their coordination number with the surface  , Similar results of a second order kinetics was reported by  for the adsorption of copper from industrial waste waters by periwinkle shells activated carbon.
3.4. Adsorption Isotherm Studies
The plots of the Langmuir, Freundlich and Temkinisotherm models for adsorption data from adsorption experiment at equilibrium pH and a temperature of 30˚C are shown in Figures 7-9 respectively. The adsorption constants of these models determined from the slopes and intercepts of these figures along with the corresponding correlation coefficients are listed in Table 3. Similar plots of the Langmuir, Freundlich and Temkin isotherms models were made for adsorption experiments at temperatures of 40˚C, 50˚C and 60˚C from where the adsorption constants and the corresponding correlation coefficients were determined. These are also listed in Table 3.
Figure 7. Langmuir isotherms for adsorption of Fe (III) ion onto PSC.
Table 3. Isotherm model constants and correlation coefficients for adsorption of Fe3+ on PSC.
Figure 8. Freundlich isotherm for the adsorption of Fe3+ ions onto PSC.
Figure 9. Temkin isotherm for adsorption of iron (III) ions onto PSC.
Table 3 shows that the correlation co-efficient (R2) of the Langmuir isotherm was closer to unity when compared to that of the Freundlich and Temkin isotherms at all temperatures the adsorption experiments were performed. Hence the Langmuir isothermf its the experimental equilibrium adsorption data of Fe3+ ions better than the Freundlich and Temkin isotherms, thus the Langmuir isotherm was most suitable for the data, followed by Freundlich then the Temkin isotherm. This suggest that the adsorption of Fe3+ ions on periwinkle shell activated carbon is a mono layer adsorption process on a homogeneous activated carbon surface. This agrees with the pseudo second order fit of the adsorption kinetics results with chemisorption (metal ions stick to the adsorbent surface) as the rate limiting step. The effect of the isotherm shape is seen in the values of the dimensionless separation factor ( ). Table 3 shows that values were between 0 and 1 at all temperatures, indicating that the adsorption process is favorable [favorable ( )]. Table 3 also shows that the maximum adsorption capacity ( ) increases with temperature, confirming that adsorption of Fe3+ ions on periwinkle shell activated carbon increases with temperature.
The relationship between and is shown in Figure 10.
Figure 10. Thermodynamic study for Fe3+ ions adsorption on PSC.
Table 4. Thermodynamic parameters for the adsorption of Fe3+ onto Periwinkle shell carbon (PSC).
The Gibb’s free energy change of the process ( ) were all negative and decreased with increase in temperature at all temperatures the experiments were conducted. This indicates that the adsorption mechanism of periwinkle shell activated carbon toward iron (III) is spontaneous in nature and thermodynamically favorable  and more favorable at higher temperatures  . This is in agreement with the favorable adsorption process predicted by the langmuir isotherm model which fitted most the adsorption isotherm data. The enthalpy change of the process ( ) is positive, hence the process is endothermic. Therefore, adsorption capacity increase with temperature increase as predicted by the values of the maximum adsorption capacities ( ) from the langmuir isotherm model and in agreement with the Gibb’s free energies obtained. The entropy change of the process ( ) is positive. Entropy is defined as the degree of freedom or chaos of a system. The positive value of the entropy change indicates that the degree of freedom or randomness at the solid/liquid interface of the activated carbon increases during the adsorption of Fe3+ ions onto the active sites of the periwinkle shell activated carbon  . This is in agreement with the work of Saha & Chowdhury,  and explained to be because the mobility of adsorbate ions/molecules in the solution increase with increase in temperature and that the affinity of adsorbate on the adsorbent is higher at high temperatures.
The adsorption performance of periwinkle shell activated carbon for the removal of iron (III) oxide from aqueous solution was investigated. The periwinkle shells were carbonized at a temperature of 300˚C and activated using nitric acid. The produced activated carbon was characterized by determining its bulk density, pH level, porosity, carbon yield, moisture content and ash content. Batch experiments were carried out to investigate the effect of some process parameters such as: contact time, particle size, adsorbent dosage and the pH on adsorption rate. The adsorption rate increased with increase in contact time, adsorbent dosage and pH of aqueous solution (reduced with increase in acidity) and decrease with increase in particle size. The kinetics of Fe3+ adsorption onto periwinkle shell activated carbon obeys the pseudo-second-order model, suggesting that chemisorption is the rate-limiting step in the adsorption process. The adsorption equilibrium data fitted the Langmuir isotherm model in comparison to the Freundlich and Temkin isotherm models at the studied temperatures, therefore periwinkle shell activated carbon had a homogeneous surface with adsorption capacity increasing with increasing temperature. The dimensionless separation factor or equilibrium parameter ( ) values calculated were greater than zero but less than 1, which showed that the adsorption process was favorable (Periwinkle shell activated carbon was an efficient adsorbent for Fe3+ ion removal). The thermodynamic parameters obtained indicated that the adsorption process of Fe+ ions was spontaneous (negative values of ) and endothermic in nature (positive ). The positive value of the entropy change ( ) showed increased randomness with adsorption. Hence the adsorption process is thermodynamically favorable.