Equilibrium studies on the uptake of nitrate and phosphate ion onto low-cost adsorbent prepared via radiation-induced graft polymerization and hydrazine hydrate functionalization

In the study, cellulosic fabric waste-based anion exchanger (‘Cell-AE’), with abundant N+(CH3)2 functional groups were prepared by graft copolymerization of acrylonitrile (AN) onto cotton fabric waste using γ radiation 60Co, followed by chemical modification with hydrazine hydrate and alkylation with dimethyl sulfate. Factors affecting the grafting process, such as radiation dose and monomer concentration, was investigated. The main adsorbent (‘Cell-AE’) and its intermediate precursors were characterized using Fourier transform infrared spectroscopy (FTIR) and scan electron microscopy (SEM). The nitrate and phosphate sorption potentials of the Cell-AE further evaluated via batch mode. Based on the results obtained, ‘Cell-AE’ showed higher adsorption affinity towards phosphate ion (19.56 mg/g), when compared to that of the nitrate ion (11.23 mg/g). Similarly, the phosphate and nitrate ion


Introduction
Freshwater, including rivers, lakes and groundwater represents about 2.66 % of the whole global water resources, of which only about 0.6 % is suitable for drinking [1]. As the global population increases, the global demand for freshwater increases and therefore, it is top necessary to treated water reserves and wastewater carefully to overcome the increasing water demand [2]. Chemical compounds composed of phosphate or nitrate group are among the leading freshwater pollutants. They find their ways into the aquatic environment, mostly due to the unwholesome wastewater discharge habits of process industries and could significantly upset the natural biological balance of living organisms present and affecting the water quality [3,4]. Water eutrophication is a form of water pollution of global concerns [5]. Although nitrate (NO3 − ) and phosphate (PO4 3− ) are important nutrients for the flourishing of plants and many other unicellular organisms, they still rank among the prominent nutrients responsible for this freshwater eutrophication [5].
The issue of eutrophication subsists when nitrate and phosphate laden waste streams from varying sources (such as, fertilizer production plant, septic tanks, atmospheric deposition, etc) is discharged into the aquatic environment in concentrations that exceeds the relevant permissible limits [6]. The presence of a high concentration of these pollutants in watercourses could ultimately cascade into several harmful ecological effects [7,8]. A significant reduction of oxygen level due to eutrophication can be very detrimental to aquatic life in general and results in reduced biodiversity [9]. Also, excessive ingestion of nitrate (NO3 − ) ion contaminated water often results in blue baby syndrome, a disease caused due to nitrate binding with haemoglobin. Phosphate and nitrate pollution can also have a severe effect on the renewability of natural resource [10][11][12]. Hence, WHO stipulated permissible drinking water limits of 40 and < 0.5 mg L −1 for NO3 − and PO4 3− contamination, respectively [13].
Considering the aforementioned negative implications of phosphate and nitrate pollution, their removal from the waste stream is imperative. To this end, several phosphate and nitrate removal techniques such as chemical precipitation [14], adsorption [15], reverse osmosis [16], biological removal [17], electrodialysis [18] have been investigated. Furthermore, some of these techniques are technologically driven and expensive in the context of developing countries. For instance, enhanced biological treatment has depicted up to 97 % phosphate removal [19], but the variation in the effluent chemical composition and temperature limits their full-scale implementation. Similarly, the high operational cost and process inefficiency associated with the application of electrodialysis and reverse osmosis has also been highlighted [19]. However, the adsorption technique has been commonly applied due to its operational flexibility and economics, as well as low sludge generation [20,21]. Also, the spent adsorbent could have agricultural application as a phosphate fertilizer and soil conditioner [22].
Native cellulosic fibres, one of the most abundant renewable raw material, with great industrial applications. However, their lack of thermo-plasticity, poor dimensional stability and poor crease-resistance retard their adsorption performance properties [23]. These properties can be improved by subjecting the cellulosic fabric to different chemical modifications for example grafting of different monomers to the cellulosic chain or etherification reaction [24][25][26][27][28][29][30][31][32]. Other performance properties, like thermal stability and resistance to biological and chemical agents, can also be improved through chemical modification and consequently applied as an absorbent in wastewater purification [33-38], ion-exchangers for the removal of metal cations from aqueous solutions [39][40][41][42][43][44], etc. Weakly basic anion exchangers contain primary, secondary, tertiary or quaternary amino groups in their chemical structures [45][46][47].
Meanwhile, the basicity of the resins containing these amino groups can be enhanced by  [58,59]. Therefore, the current study aims at grafting acrylonitrile (AN) onto cotton fabric, followed by reaction with hydrazine hydrate and finally alkylation with dimethyl sulfate. The synthesized adsorbent (Cell-AE) was further applied for removing phosphate and nitrate anions from their aqueous solutions. The effect of the process variable on the anion removal was investigated, while the adsorbent properties were characterized using Fourier transform infrared spectroscopy (FTIR) and scanning electron microscopy (SEM).

Synthesis of adsorbent
Poly-(acrylonitrile)-grafted cellulose was prepared via radiation-induced polymerization reaction. The grafting reaction was carried out in a pyrex tube containing the monomer dissolved in DMF, together with a definite weight of the cellulosic fabric and 5 % styrene (based on monomer weight) as an inhibitor. The reaction was carried out using γ-rays of 60 Co source at different radiation doses. The grafted cotton fabric was removed from the Pyrex tube, washed thoroughly with DMF, then several times with distilled water to remove the homopolymer from the grafted cotton fabric, and then the grafted fabric was dried and weighed.
The degree of grafting was calculated as a percentage increase in the fabric weight following Where, Co and Ce (mg L −1 ) are the initial and final anion concentrations, respectively. V (L) is anion solution volume and W (g) is the adsorbent fabric weight. For accuracy, the adsorption experiments were performed always in duplicate to get more accurate mean values for qe.

Instrumental characterizations
The untreated, grafted and modified cotton fabrics were examined by Fourier Transform infrared spectroscopy (FTIR) to assign the vibrational frequencies of different functional groups present in their structures. The IR spectra of the samples were recorded using Mattson 1000 FTIR spectrophotometer (Unicom, England).
The scanning electron micrograph (SEM) test was conducted to characterize the morphology and surface properties of untreated, grafted and modified cotton fabric. The samples were mounted on a standard microscope stub and coated with a thin gold layer by use of a JEOL SEM 25 (Japan). 6
The nonlinear model goodness of data fit is usually evaluated from the magnitude of some  Table 2.

Model name Model Equation Model name Model Equation
Langmuir

Adsorbent synthesis mechanism
The suggested mechanism for the reaction of cotton fabric with acrylonitrile (AN), followed by hydrazine hydrate functionalization and final alkylation using dimethyl sulfate is discussed  This linear relationship subsequently levelled off at higher monomer concentrations. This variation in the grafting rate may be explained in terms of the non-radical grafting mechanism, where for this type of heterogeneous polymerization reaction represented in the study, the monomer is somehow distributed between the aqueous phase and the stationary cellulose interface. Meanwhile, the relative extent of such distribution is dependent on the polarity of the monomer molecules. It is also expected that a relatively high proportion of a polar monomer, such as acrylonitrile, will be in the aqueous phase and accordingly, a reduced amount of acrylonitrile molecules will be available at the cellulose interface (grafting site) and hence low grafting levels are expected [85].
Using the procedure of styrene comonomer [86], it was observed that the involvement of styrene as a component of the monomer mixture, enhanced the copolymerization of the second reactive monomer via mutual irradiation with a minimum formation of homopolymer. The utilization of such a technique leads to a remarkable reduction in the acrylonitrile key performance index (Kp) in the presence of styrene. In other words, the presence of 0.05% (w/v) of styrene in the monomer feed, results in the reduction of acrylonitrile homo-polymerization reaction and accordingly decreases its competition with the desired graft polymerization reaction. Figure 1 also depicts the relationship between the graft rate variation with varying radiation doses (5KGy to 15 KGy). The plot shows an initial increase in the extent of grafting when the radiation doses increased up to 10 KGy. It was however observed that the grafting rate also tends to level off when the radiation dose increased beyond 10 KGy. The number of free radicals created by the radiation process initially increased linearly with increasing radiation dose. Meanwhile, at higher radiation doses (beyond 10 KGy), the number of free radicals decreased, due to the recombination of the formed radicals, useful for initiating the graft polymerization reaction [87] Also, at this levelling off point, the grafting reaction itself becomes a diffusion-controlled process [88].    Meanwhile, the effect of variation of adsorbate concentration on the equilibrium adsorption data was fitted to the isotherm models whose equations are depicted in Table 1, while the associated parametric and R 2 -values are presented in Tables 3 and 4 for nitrate and phosphate adsorption, respectively. The graphical illustration of the experimental dataset for nitrate and phosphate adsorption was plotted and presented in Figure 4 and 5, respectively. Generally, certain isotherm model parameters provide reasonable information on the nature of the adsorption system. For instance, when the Langmuir separation factor, (RL) is either greater than or equal to unity, the adsorption process is considered to be linear and unfavourable, respectively. Similarly, when the RL value is greater than zero but less than unity, a favourable adsorption system is implied, while, irreversible adsorption is signified by RL= 0. Furthermore,  Tables 3 and 4 for both anions indicates the exothermicity of the respective anion adsorption onto 'Cell-AE'.

SEM analyses of the samples
Consequently, the calculated D-R mean adsorption energy, E (kJ/mol) for both anions which are greater than 10 kJ/mol, favoured the occurrence of a chemical adsorption process.
Meanwhile, the insufficiency of the application R 2 -value alone for the determination of the best fit nonlinear model has been highlighted [2]. As a result, seven (7) error model, whose equations are presented in Table 2 were applied for determining the best-fit isotherm model.
To limit the inconsistencies often experienced during the application of multiple error model (as was the case in this study), a process of normalizing the different error values from the error models for a given isotherm was adopted. Consequently, only the sum of normalized error (SNE) value will be considered during the isotherm modelling discussion and the lower the SNE value (as shown in Tables 5 and 6), the better the model fit to the experimental isotherm data.
The Redlich-Peterson (R-P) and Dubinin-Radushkevich (D-R) models, respectively emerged as the overall best fit for nitrate and phosphate ions, since both models returned the lowest SNE value (Tables 5 and 6), together with an appreciably high R 2 -value (Tables 3 and 4).