
Research Article


Nitrate Adsorption by PanOximeNano Fe_{2}O_{3} Using a TwoLevel Full Factorial Design 

Ramin Nabizadeh
and
Mahsa Jahangiri Rad



ABSTRACT

PANoximenano Fe_{2}O_{3} prepared and used as adsorbents for the adsorption studies of nitrate in aqueous solution. The main and interactive effects of two various experimentally controlled environmental factors namely, initial nitrate concentration and time of reaction were investigated through the model equations designed by a twolevel full factorial design in a shakeflask system. The results predicted using factorial regression model showed high values of regression coefficients (R^{2} = 0.959) showing good agreement with experimental data. Main and interacting effects of different process parameters on the response (amount of nitrate removal adsorbed) were found. The maximum adsorption of 24 mg g^{–1} achieved at initial concentration of 200 mg L^{–1} and reaction time of 50 min.





Received: July 25, 2015;
Accepted: September 18, 2015;
Published: December 15, 2015


INTRODUCTION
There has long been concern expressed with the presence of nitrate in drinking water at concentrations exceeding the World Health Organization guideline of 50 mg L^{–1} (WHO., 2006). Nitrogen species, like nitrate and nitrite, have been confirmed to be procarcinogenic. The excessive input of these nitrogen compounds into food and drinking water in the presence of tertiary amines, can lead to the formation of carcinogenic Nnitrosoamines (Romano and Zeng, 2009). The nitrate content in crops is one of the most important indicators of farm production quality. Nitrate content in food is strongly regulated due to its toxicity, predominantly to young children. The actual toxin is not the nitrate ion itself but rather the nitrite ion which is formed when nitrate is reduced by intestinal bacteria (Majumdar and Gupta, 2000). Nitrite ions are toxic because they can combine with hemoglobin with resulting formation of met hemoglobin. The association constant for met hemoglobin formation is larger than that for oxyhemoglobin complex formation. Thus, the nitrite ion binds with hemoglobin, depriving the tissues of oxygen. Severe cases of disease called methemoglobinemia can result in mental retardation of the infant and even a death (Tate and Arnold, 1990). Principal sources of nitrate in water are runoff and drainage from land treated with agricultural fertilizers and also deposition from the atmosphere as a consequence of NO_{x} released from fossil fuel combustion (Velizarov et al., 2004). On the other hand, nitrate is often found in the effluents of biological treatment plants because it represents the final form of nitrogen from the oxidation of organic nitrogen compounds. The nitrogen present in soil organic matter may also be released as nitrate through microbial action (Haugen et al., 2002). Moreover, nitrate’s role as a nutrient contributes significantly to blooms of algae (Masukume et al., 2013). The major source of drinking water from private water wells located near each private residence in the rural communities, often exceed the US EPA maximum contaminant level for drinking water of l0 mg L^{–1} (NO_{3}N) (Prakasa Rao and Puttanna, 2000). Isotopic investigations, based upon ^{15}N/^{14}N, have revealed the predominance of anthropogenic, organic human and animal waste and decay of irrigation enhanced vegetation are major causes of nitrate contaminated water rather than natural nitrate sources (Bhatnagar and Sillanpaa, 2011). Several adsorbents have been applied to remove nitrate from drinking water (Nabizadeh et al., 2015). In our prior studies, the synthesis of PANoximenano Fe_{2}O_{3} and its possible application as an adsorbent for nitrate and fluoride removal were presented (Nouri et al., 2014). Several factors affect adsorption including contact time and initial concentration. This study set out with the aim of assessing the importance of variables as well as searching optimum conditions of variable to predict targeted responses (nitrate adsorption). Response Surface Methodology (RSM) has been proven to be successfully implemented in some researches, such as engineering, food processing, biotechnology and adsorption processes for optimization. Thus Optimization of the process parameters affecting the nitrate adsorption process was made by RSM under "R" software (Lenth, 2009).
MATERIALS AND METHODS Preparation of PANoximenano Fe_{2}O_{3}: The preparation and synthesis of selected PANoximenano Fe_{2}O_{3} was earlier described in our previous research (Lenth, 2009). Figure 1 shows the hypothesized molecular reaction of functionalized polyacrylonitrile coated with iron oxide nanoarticles. Pulse Electrical Explosion (PEE) Nano Engineering and Manufacturing Co. (Model: PEE10k or PEE50K) was used to characterize nanoparticles properties.
 Fig. 1:  Expanded molecular formula of functionalized PAN coated with iron oxide nanoparticles 
Batch study: Stock solution of nitrate (1000 mg L^{–1}) was prepared by dissolving sodium nitrate in distilled water and certain concentrations of nitrate were obtained by further dilutions. The pH of each last solution was adjusted to the natural pH value of drinking water (pH = 7) with diluted and concentrated H_{2}SO_{4} and NaOH solution. All the chemicals used were of analytical reagent grade. Batch experiment was carried out with variation of the two factors including initial concentration of nitrate aqueous solution and contact time. All batch adsorption experiments were carried out in 50 mL sealed plastic tubes with the working volume of 25 mL. After adding a known weight of the adsorbent (30 mg), the flask was shaken (150 rpm) for 24 h in a temperature controlled (25±2°C) water bath shaker for predetermined time intervals. The liquid samples then centrifuged and residual fluoride was analyzed by use of spectrophotometer. The amount of nitrate uptake adsorbed was calculated from the Eq. 1: where, q_{e} is the nitrate adsorbed (mg g^{–1}), C_{o} and C_{e} are the initial and equilibrium concentrations of nitrate in solution (mg L^{–1}), respectively, V is the volume of solution and m is mass of the adsorbent (g). A second order polynominal model with interaction terms fitted to the experimental data obtained from the experimental runs was conducted by "R" software (Kiran et al., 2007). Modeling by full factorial design: Since a number of factors such as initial concentration and time of reaction influencing the adsorption process, these parameters were studied. The studying of the each and every factor is quite tedious and time consuming. Thus, a factorial design can minimize the above difficulties by optimizing all the affecting parameters collectively at a time (Amini et al., 2008). Factorial design comprises the great precision in estimating the overall main factor effects and interactions of different factors. In full factorial design every setting of factor appears with every setting of every other factor. Factorial designs are strong candidates in examining treatment variations. Instead of conducting a series of independent studies we can combine these studies into one. A common experimental design is one with all input factors set at twolevels each (Perry et al., 1997). These levels are called ‘High’ and ‘Low’. The levels and ranges of the studied process parameters (X_{1}time) and (X_{2}concentration). Table 1 shows the experimental ranges of each factor.
Table 1:  Experimental ranges and levels of the factors used in the factorial design by "R" software 
 
Statistical analysis: In order to determine whether or not the second order polynominal equation was significant to fit with the experimental result an analysis of variance (ANOVA) was conducted using "R" software. RESULTS AND DISCUSSION Adsorbent characteristic: Figure 2 represents the XRD patterns functionalized PANFe_{2}O_{3}. The XRD patterns of functionalized PANFe_{2}O_{3} (Fig. 2) show a broad noncrystalline peak (2θ = 2030°) and a crystalline peak (2θ = 18°) corresponding to the orthorhombic PAN reflection and besides the diffraction peaks of the PAN phase, another peaks appeared corresponding to the above peaks, indicating that the Fe_{2}O_{3} nanoparticles on the PAN have the same crystal diffraction as pure Fe_{2}O_{3}. The TEM images of as prepared nanomaterial presented in Fig. 3 is the image of PANoximeFe_{2}O_{3}, showing that Fe_{2}O_{3} nanoparticles are well dispersed with an average diameter of around 50 nm. Main and interaction effects: The application of the response surface methodology based on the calculation of parameters showed an empirical relationship between the response and the input variables projected by the following fitted second order polynomial equation:
The student tdistribution and the corresponding pvalues, along with the parameter estimates, are listed in Table 2. The data obtained from Eq. 2 is significant. Significance of each coefficient of Eq. 2 was determined by applying ttest and pvalues. Since, the pvalues of the all the coefficients, are p<0.05 it implies that these are significant. A larger t value and smaller pvalue identifies the effect that appears to be very prominent (Elibol, 2002; Gurses et al., 2002). According to Eq. 2, it is clear that the first and second order main effects of all the two variables, namely initial nitrate concentration and time of reaction are significant due to their corresponding low p value.
 Fig. 2:  XRD pattern of PANoximenanoFe_{2}O_{3} 
 Fig. 3:  TEM image of PANoximeFe_{2}O_{3} 
Table 2:  Estimated parameters and their significance 
 
The quantities X_{1}, X_{2} exhibited positive influence, whereas the coefficients of the quadratic terms had negative effects. on adsorption. Based on above equation the effect of time is more influential than time of adsorption due to the higher coefficient (0.63).
Statistical analysis: In order to determine whether or not the second order polynomial equation was significant to fit with the experimental result, it is necessary to conduct an analysis of variance (ANOVA) (Myers and Montgomery, 1995). The results of ANOVA for the second order equation is presented in Table 3. Based on the ANOVA analysis the equation represents adequately the actual relationship between the response (the amount of nitrate adsorbed (mg) on adsorbent (g)) and the significant variables. The coefficient of determination (R^{2}) is found to be 0.959 which is very high and shows high correlation between the observed and the predicted value. The Fisher Ftest with a very low probability value (p), also demonstrates a very high significance for the second order equation.
Response surface counter plot: Response surface contour plots help to understand the relationship between the response and experimental levels of each variable. These plots also show the type of interaction between test variables and help to obtain the optimum conditions (Box et al., 1978). Contour plots for the significant interactions of time and concentration on nitrate adsorption are shown in Fig. 4. As depicted, both time and initial nitrate concentration played major role and in nitrate uptake.
Table 3:  Analysis of variance for fitted second order polynomial equation 
 Residual standard error: 1.38 on 24 degrees of freedom. Multiple Rsquared: 0.9599, adjusted Rsquared: 0.9515, df: Degree of freedom 
 Fig. 4:  Response surface contour plots showing effect of initial nitrate concentration and time on nitrate removal 
The removal uptake of nitrate is high at higher temperature and concentration than at their lower values. It can be also observed from the counter plot that at constant concentration as temperature rises, the nitrate uptake increase profoundly. Maximum nitrate uptake was observed at higher concentration and temperature (24 mg g^{–1}). A circular contour plot represents that the interaction between the corresponding variables is minimal while elliptical contour plot indicates significant interactions (Antony, 2003). Elliptical contour plot obtained from the data obviously shows that the mutual interactions between the variables are significant.
Normal probability plot of residuals: The normality of the data can be checked by plotting a normal probability plot of the residuals. If the data points on the plot fall fairly close to the straight line, then the data are normally distributed (Antony, 2003). The normal probability plot of the residuals is shown in Fig. 5. It can be seen that for the data points were fairly close to the line and it indicates that the experiments come from a normally distributed population.
Time concentration effect plot: Interactions of initial nitrate concentration and time were interpreted with effect plot under "R" software (Fig. 6). Their interactions can be inspected by restricting the linear model to these two factors with their interactions.

Fig. 5(ab): 
Normal probability plots of residuals 
As moving from left side to the right side of the plot the nitrate adsorption increases due to the increase in concentration and time of adsorption.
 Fig. 6:  Interaction plots for the nitrate adsorption experiment from linear model with interactions on the two factors identified before 
This plot also predicts nitrate adsorption when time of reaction move to the upper lower limit of experiment. As depicted at maximum initial itrate concentration (200 mg L^{–1}) when time of adsorption goes beyond 50 min, adsorption equilibrium reached 45 mg g^{–1}.
CONCLUSION PANoxime nano Fe_{2}O_{3} was used as an adsorbents for nitrate removal. The effects of two parameters affecting adsorption namely, initial nitrate concentration and time of adsorption were investigated at constant pH of natural water condition (pH = 7±0.5). Analysis of results carried out by using twolevel full factorial design. The response surface methodology modal chosen for our study showed high correlation coefficient (0.959). The Ftest with a very low probability value (p), also demonstrates a very high significance for the second order equation.

REFERENCES 
WHO., 2006. Guidelines for DrinkingWater Quality: Incorporating First Addendum Recommendations. Vol. 1, 3rd Edn., World Health Organization, Geneva, Switzerland, pp: 375376
Romano, N. and C. Zeng, 2009. Evaluating the newly proposed protocol of incorporated potassium in nitrate toxicity experiments at different salinities: A case study with the tiger prawn, Penaeus monodon, juveniles. Aquaculture, 289: 304309. CrossRef  Direct Link 
Majumdar, D. and N. Gupta, 2000. Nitrate pollution of groundwater and associated human health disorders. Indian J. Environ. Health, 42: 2839. Direct Link 
Tate, C.H. and K.F. Arnold, 1990. Health and Aesthetic Aspects of Water Quality. In: Water Quality and Treatment, Pontius, F.W. (Ed.). McGrawHill Inc., New York, pp: 63156
Velizarov, S., J.G. Crespo and M.A. Reis, 2004. Removal of inorganic anions from drinking water supplies by membrane bio/processes. Rev. Environ. Sci. Bio/Technol., 3: 361380. CrossRef  Direct Link 
Haugen, K.S., M.J. Semmens and P.J. Novak, 2002. A novel in situ technology for the treatment of nitrate contaminated groundwater. Water Res., 36: 34973506. CrossRef  Direct Link 
Masukume, M., M.S. Onyango, O. Aoyi and F. Otieno, 2013. Nitrate removal from groundwater using modified natural zeolite. http://www.ewisa.co.za/literature/files/144_97%20Masukume.pdf.
Prakasa Rao, E.V.S. and K. Puttanna, 2000. Nitrates, agriculture and environment. Curr. Sci., 79: 11631169. Direct Link 
Bhatnagar, A. and M. Sillanpaa, 2011. A review of emerging adsorbents for nitrate removal from water. Chem. Eng. J., 168: 493504. CrossRef  Direct Link 
Nabizadeh, R., M. JahangiriRad, M. Yunesian, J. Nouri, F. Moattar and F. Sadjadi, 2015. Synthesis and characterization of functionalized polyacrylonitrile coated with iron oxide nanoparticles and its applicability in nitrate removal from aqueous Solution. Desalin. Water Treat. CrossRef  Direct Link 
Nouri, J., R. Nabizadeh, M. Jahangirirad, M. Yunesian and F. Moattar, 2014. Fluoride removal from aqueous solution by functionalizedpolyacrylonitrile coated with iron oxide nano particles: characterization and sorption studies. Desalin. Water Treat., 52: 43694375. CrossRef  Direct Link 
Lenth, R.V., 2009. Responsesurface methods in R, using RSM. J. Stat. Software, Vol. 32, No. 7. CrossRef  Direct Link 
Kiran, B., A. Kaushik and C.P. Kaushik, 2007. Response surface methodological approach for optimizing removal of Cr (VI) from aqueous solution using immobilized cyanobacterium. Chem. Eng. J., 126: 147153. CrossRef  Direct Link 
Amini, M., H. Younesi, N. Bahramifar, A.A.Z. Lorestani, F. Ghorbani, A. Daneshi and M. Sharifzadeh, 2008. Application of response surface methodology for optimization of lead biosorption in an aqueous solution by Aspergillus niger. J. Hazard. Mater., 154: 694702. CrossRef  Direct Link 
Perry, R.H., D.W. Green and J.O. Maloney, 1997. Chemical Engineers Handbook. 7th Edn., McGrawHill, New York
Elibol, M., 2002. Response surface methodological approach for inclusion of perfluorocarbon in actinorhodin fermentation medium. Process Biochem., 38: 667673. CrossRef  Direct Link 
Gurses, A., M. Yalcin and C. Dogar, 2002. Electrocoagulation of some reactive dyes: A statistical investigation of some electrochemical variables. Waste Manage., 22: 491499. CrossRef  Direct Link 
Myers, R.H. and D.C. Montgomery, 1995. Response Surface Methodology: Process and Product Optimization Using Designed Experiments. 1st Edn., John Wiley and Sons, New York, USA., ISBN13: 9780471581000, Pages: 700
Box, G.E.P., W.G. Hunter and J.S. Hunter, 1978. Statistics for Experimenters. An Introduction to Design, Data Analysis and Model Building. John Wiley and Sons Inc., New York, USA., Pages: 678
Antony, J., 2003. Design of Experiments for Engineers and Scientists. ButterworthHeinemann Publisher, Jordan Hill, Oxford, ISBN10: 0750647094, Pages: 176



