Tuning N-methyl-D-glucamine density in a new radiation grafted poly(vinyl benzyl chloride)/nylon-6 fibrous boron-selective adsorbent using the response surface method

Abstract

A new adsorbent containing boron-selective groups was prepared by radiation induced grafting of vinyl benzyl chloride (VBC) onto nylon-6 fibers followed by functionalisation with N-methyl-D-glucamine (NMDG). The density of glucamine loaded in the adsorbent was tuned by optimisation of the reaction parameters such as NMDG concentration, reaction temperature, reaction time and degree of grafting using response surface methodology (RSM) employing Box–Behnken design (BBD). The optimum parameters for achieving the maximum glucamine density (1.7 mmol g⁻¹) in the adsorbent are 10.6%, 81 °C, 47 min and 121% for the NMDG concentration, reaction temperature, reaction time and degree of grafting (DG), respectively. The deviation between the optimum experimental and predicted glucamine density is found to be 1.2% suggesting the reliability of RSM in predicting the yield and optimising the functionalisation reaction parameters. The chemical composition, morphology and structure of the NMDG-containing fibrous adsorbent were studied using Fourier-transform infrared (FT-IR) spectroscopy, scanning electron microscopy (SEM) and X-ray diffraction (XRD), respectively. The thermal properties were determined using differential scanning calorimetry (DSC) and the thermal stability was evaluated by thermogravimetric analysis (TGA). Considering the physico-chemical properties of the fibrous adsorbent and the preliminary results of boron adsorption, it can be suggested that this adsorbent is a promising candidate for boron removal.

---

1. Introduction

The increase in the level of boron pollution found in surface and wastewaters has generated widespread attention in recent years to find an effective and economically-viable boron removal method. This arose from the constant growing demand for applications of boron and its compounds in various industries including metallurgy, medicine, glass, ceramics, detergents, fertilizers and insecticides. Moreover, naturally occurring boron present in drinking water and geothermal water has to be treated to low levels to meet the stagnant regulations set by various governmental and environmental agencies and authorities across the world.¹

Various technologies have been used to remove boron from different streams including precipitation–coagulation, chemical adsorption, physical adsorption, electro-dialysis, liquid–liquid extraction, reverse osmosis, electro-coagulation and ion exchange.² Of these, ion exchange is the most effective and preferable method for removing boron to meet the required levels compared to other separation methods.³ Chelating/ion exchange resins represent the central component in ion exchange systems and their distinct affinity towards boron is crucial for the viability of ion exchange treatment of boron-contaminated streams.

Commercial boron chelating resins and their researched counterparts based on functionalised polymers mostly contain N-methyl-D-glucamine (NMDG). This functional group exhibits a high selectivity towards boron due to the presence of five hydroxyl groups allowing for the coordination of boron to form a stable complex.⁴ Therefore, the majority of researched adsorbents are obtained by the functionalisation of adsorbent precursors with NMDG.³

Various polymer-based adsorbents have been developed for boron removal from solution using various polymerisation routes.^5–12 A review of various strategies for the preparation of alternative polymer-based adsorbents and their performance evaluation was published recently.³ Among all, adsorbents obtained by radiation-induced grafting and subsequent functionalisation provide attractive materials for more effective removal of ionic pollutants from water and wastewater.¹³ Radiation-induced grafting is used for the preparation of such desired materials due to its simplicity and as it does not require catalysts or initiators, thus leaves no detrimental residues. Moreover, the grafting reaction can be initiated from monomers in different phases and temperatures allowing either surface or bulk modification of polymer substrates depending on the reaction parameters which can be tuned to obtain the desired compositions. Various polymer substrates such as fibers, fabrics and films can be modified with a variety of functional groups in large quantities.^13,14 Particularly, radiation grafted adsorbents having a fibrous structure with high selectivity and a large surface area overcome the diffusion limitations associated with the conventional granular resins and showed a promising performance for the removal of pollutants such as heavy metal ions¹⁵ and phosphate as well as arsenic species.¹⁶

Removal of boron by fibrous adsorbents grafted with glycidylmethacrylate (GMA) onto various fibers such as polypropylene and viscose fibers,⁹ nylon-6 fibers,¹⁷ polyethylene (PE)¹⁸ fibers and PE non-woven fabric¹⁹ followed by NMDG treatment was reported recently. The fibrous chelating adsorbents showed a higher adsorption capacity and faster kinetics together with a tolerance to high flow speed with a small drop in the breakthrough capacity compared to chelating granular resins.^8,9,19 However, these adsorbents are challenged by degradation (leaching of functional groups) during the regeneration process in addition to deterioration of their mechanical properties caused by the incorporation of excessive amounts of amorphous poly(GMA) during grafting to impart a high glucamine content leading to a fragile structure.

To improve the chemical and mechanical stability of the fibrous adsorbents, vinyl-benzyl chloride (VBC) monomer has been grafted as an alternative monomer to confer more stable benzyl arms to a favourable fibrous substrate (nylon-6 fibers) capable of hosting glucamine through a facile functionalisation reaction. VBC has neither been radiation grafted on nylon-6 fiber nor used for the preparation of fibrous adsorbents for boron removal. The selection of nylon-6 as a substrate is due to its outstanding physico-mechanical properties and low cost.²⁰ Moreover, it has a moderate radiation resistance where it undergoes crosslinking and chain scission. The extent of each radiolytic reaction depends on the adsorbed dose and irradiation atmosphere.²¹ Crosslinking, which takes place in nylon-6 during irradiation, is reported to improve the mechanical properties compared to original nylon-6.²⁰ Thus, it is of high interest to tune the content of the glucamine groups to minimize the impact of the applied preparation procedure (grafting and functionalisation) on the properties of the nylon-6 fibrous substrate. This can be achieved by modelling the functionalisation reaction, optimisation of its parameters and prediction of the reaction yield using statistical tools.

The response surface method (RSM) developed by Box and Wilson²² is a mathematical expression to describe the relationship of a few independent variables with one or more responses. RSM is based on fitting mathematical models to the experimental results generated from a designed experiment to verify the model obtained. When the model is identified, optimisation and predictions of the experimental parameters such as variables and responses can be derived according to requirements. These include improving the process performance, identifying significant relationships of parameters, reducing operational costs and minimizing the experimental time.²³ The developed experimental design is based on three-level (+1, 0, −1) incomplete factorial designs which take the interactions of variables into account.²⁴

The objective of this study was to prepare a new boron-selective adsorbent with a tuned content of glucamine groups. This was carried out by radiation-induced grafting of VBC onto nylon-6 fibers followed by functionalisation with NMDG under optimised parameters determined through the RSM. The properties of the prepared NMDG-containing fibrous adsorbent were evaluated using scanning electron microscopy (SEM), Fourier transform infrared (FT-IR) spectroscopy, differential scanning calorimetry (DSC), thermogravimetric analysis (TGA), X-ray diffraction (XRD) and a universal mechanical tester with reference to the original and poly(VBC)-grafted nylon-6 fibers. The obtained adsorbent was tested for the removal of boron from solutions for various pH values and adsorption/desorption cycles.

2. Experimental

2.1 Materials

VBC (a mixture of 3- and 4-isomers, 97% containing 700–1100 ppm nitromethane and 50–100 ppm tert-butylcatechol inhibitors) and NMDG were purchased from Sigma-Aldrich. Solvents such as 1,4-dioxane and methanol were supplied by JT Barkers. Boric acid (H₃BO₃), HCl and NaOH were supplied by Merck. Nylon-6 fibers (15 μm diameter) with a molecular weight in the range of 1.59–2.84 × 10⁶ and a moisture content of 3.5% were supplied by Reliance Sdn. Bhd. (Malaysia). Deionized water (18 mΩ cm) obtained from a water deionizer (Millipore Direct-Q™) was used to prepare the boron solutions. All the chemicals purchased were of analytical grade and used as received without further treatment.

2.2 Preparation of the adsorbent

2.2.1 Grafting of VBC onto nylon-6 fibers. Clean nylon-6 fibers of known weight were placed in a PE bag, purged with N₂ gas to remove air and thermally sealed. The sample bags were placed on a conveyor tray and irradiated to 300 kGy using an electron beam (EB) accelerator (EPS 3000, Nissin High Voltage, Japan) at 10 kGy per pass. The irradiated samples were placed in ampoules containing VBC solution diluted with methanol, and the mixture was deaerated by flushing with N₂ gas for 30 min. The concentration of the VBC monomer was varied in the range of 5–20 wt% to obtain samples with different degrees of grafting. The reaction was allowed to proceed for 3 h at 30 °C. After the reaction completed, the grafted nylon-6 fibers were removed and thoroughly washed using methanol several times. Subsequently, the grafted fibers were dried in a vacuum oven at 40 °C for 20 h before weighing.

The degree of grafting (DG) was calculated from the weight increase by:

where W_i and W_f are the weights of the original and grafted nylon-6 fibers, respectively.

2.2.2 Functionalisation with NMDG. An NMDG solution prepared by mixing an amount of NMDG in the range of 5–15 wt% with 1,4-dioxane was used to introduce glucamine to the poly(VBC) grafted nylon-6 fibers having DG in the range of 70–130% in a glass reactor under reflux. The reaction temperature was varied in the range of 70–90 °C. The optimum parameters for chemical modification were determined according to the RSM design described next. At the end of the reaction, the functionalised fibers were removed, washed thoroughly with deionized water and dried under vacuum at 40 °C for 20 h.

Fig. 1 shows a scheme for the grafting of VBC onto nylon-6 fibers and the functionalisation of poly(VBC) grafted fibers with NMDG. The density of glucamine groups loaded on the adsorbent was calculated by:

where Z_i and Z_f are the weights of the grafted fibers before and after functionalisation and M is the molecular weight of NMDG which equals 195.21.

Fig. 1 Scheme for the grafting of VBC onto nylon-6 fibers and the functionalisation of poly(VBC) grafted fibers with NMDG.

2.3 Experimental design of the response surface methodology

Box–Behnken design (BBD) was used to develop a statistical (quadratic) model for the optimisation of the independent variables for the functionalisation of the poly(VBC) grafted fibers by reaction with NMDG. The developed model can be used to evaluate the linear effects, interactions and quadratic effects of the independent variables on the response. The BBD consists of a set of points at the midpoint of each edge and the replicated center point of a multi-dimensional cube. The ranges of the variables were selected based on the preliminary experiments and the related literature of single factor experiments.

The independent variables for the present reaction include the NMDG concentration (X₁), reaction temperature (X₂), reaction time (X₃) and degree of grafting (X₄). The response or the dependent variable is the density of glucamine groups. A four factors at three levels and 29 runs BBD experimental design was applied to derive a quadratic polynomial equation to predict the optimum combination of independent variables to tune the glucamine content in the adsorbent. The 29 experimental runs of BBD composed of 16 factorial points, 8 axial points and 5 replicates at the center points. The selection ranges of the respective variable were coded as −1, 0, and +1 corresponding to the low level, mid-level and high level as shown in Table 1. The experimental ranges selected for the independent variables were: 5, 10 and 15 wt% for the NMDG concentration, 70, 80 and 90 °C for the reaction temperature, 10, 35 and 60 min for the reaction time and 70, 100 and 130% for the degree of grafting. Table 2 shows the design arrangement and the actual experimental design matrix. The Design-Expert software (Version 6.0.8, Stat-Ease, Inc., USA) was employed to analyse the experimental results.

Table 1 Code, levels and factor values for Box–Behnken design

Variables	Code	Level values	Factor values
NMDG concentration (wt%)	X1	−1, 0, 1	5, 10, 15
Reaction temperature (°C)	X2	−1, 0, 1	70, 80, 90
Treatment time (min)	X3	−1, 0, 1	10, 35, 60
DG (%)	X4	−1, 0, 1	70, 100, 130

---

Table 2 Independent variables and experiment results of the response surface analysis

Runs	NMDG (X1)	Temperature (X2)	Time (X3)	DG (X4)	Functional group density (mmol g^−1-adsorbent)
	(wt%)	(°C)	(min)	(%)	Actual	Predicted
1	5	70	35	100	1.37	1.39
2	15	70	35	100	1.51	1.53
3	5	90	35	100	1.46	1.47
4	15	90	35	100	1.60	1.61
5	10	80	10	70	0.95	0.98
6	10	80	60	70	1.41	1.40
7	10	80	10	130	1.27	1.30
8	10	80	60	130	1.70	1.69
9	5	80	35	70	1.27	1.27
10	15	80	35	70	1.40	1.40
11	5	80	35	130	1.56	1.57
12	15	80	35	130	1.70	1.71
13	10	70	10	100	1.11	1.13
14	10	90	10	100	1.16	1.18
15	10	70	60	100	1.52	1.51
16	10	90	60	100	1.63	1.62
17	5	80	10	100	1.15	1.10
18	15	80	10	100	1.30	1.24
19	5	80	60	100	1.48	1.51
20	15	80	60	100	1.63	1.65
21	10	70	35	70	1.31	1.30
22	10	90	35	70	1.35	1.34
23	10	70	35	130	1.59	1.57
24	10	90	35	130	1.70	1.68
25	10	80	35	100	1.57	1.57
26	10	80	35	100	1.57	1.57
27	10	80	35	100	1.57	1.57
28	10	80	35	100	1.57	1.57
29	10	80	35	100	1.57	1.57

---

ANOVA was used to analyse the experimental data. The coefficient of determination (R²), adjusted coefficient of determination and predicted coefficient of determination analysis were calculated to test the adequacy of the developed models. After fitting the models, the generated data were used to generate response surfaces and contour plots. All statistical analyses were conducted using Design-Expert software (Version 6.0.8, Stat-Ease, Inc., USA).

2.4 Characterization of the fibrous adsorbent

SEM images of the fibrous samples were obtained using a FEI Quanta 4000 scanning electron microscope. The samples were mounted on the sample holder and sputter-coated with a gold layer before the images were captured at a voltage of 20.0 kV.

FT-IR analysis was conducted using a PerkinElmer Spectrum One FT-IR spectrometer. The samples were scanned in the transmission mode with a resolution of 4 cm⁻¹ and wave number range of 500 to 4000 cm⁻¹.

XRD analysis was carried out using an X’pert Pro X-ray diffractometer model PW 3040 by Philips. The measurements were performed using CuKα radiation, 2θ degrees in the range of 10–80° and a counting time of 2 s. The applied voltage and current were fixed at 40 kV and 30 mA, respectively. The X-ray wavelength was fixed at 1.54 Å.

DSC measurements were performed using a Mettler-Toledo, DSC-822e calorimeter. The thermal analysis was conducted in the temperature range of 30–300 °C at a heating rate of 10 °C min⁻¹ under a N₂ atmosphere.

TGA analysis was carried out using a PerkinElmer model TGA-STA6000. The analysis was performed in the temperature range of 35–800 °C at a heating rate of 10 °C min⁻¹ and under a N₂ atmosphere.

The tensile strength and displacement were measured using an Instron universal tester (Model 4301) operated at a 10 mm min⁻¹ crosshead speed according to ASTM D3822-07. The reported values are averages of five sample measurements.

2.5 Boron adsorption and elution experiments

The adsorption of boron on the newly prepared adsorbent was tested by adding an adsorbent dose of 0.5 g to 150 mL of a 100 mg L⁻¹ solution of boric acid. The pH was varied from 3 to 11 using 1 M HCl and 1 M NaOH solutions at 30 °C. The optimum pH was determined and used in other experiments.

The effect of adsorbent dose variation on the boron removal efficiency was investigated by changing the weight in the range of 0.05–1.00 g. The experiments were carried out using 30 mL of boric acid solution with two different concentrations of 100 and 200 mg L⁻¹, a temperature of 30 °C, stirring speed of 200 rpm and reaction time of 2 h.

The elution (desorption) experiments were carried out by equilibrating the boron loaded adsorbent with a 1 M HCl solution to test the reusability of the adsorbent.

In all the adsorption experiments, the concentration of boron in the solutions was determined using a PerkinElmer Optima 7300 DV ICP-OES.

The adsorption capacity (q), which is the amount of ions adsorbed per unit mass of the adsorbent was calculated using eqn (3).

where C_i and C_f are the initial and final boron concentrations (mg L⁻¹) in the solution, respectively, W is the weight of the adsorbent used (g) and V is the volume of the solution (L). The efficiency of boron removal was calculated using eqn (4).where C_i and C_f are the initial and final boron concentrations (mg L⁻¹) in the solution, respectively.

3. Results and discussion

3.1 Response surface model fitting

The adequacy of the model’s tested output obtained from the BBD analysis indicates that the linear and quadratic models were statistically significant with p-values lower than 0.0001 as shown in Table 3. The cubic model was found to be aliased and therefore could not be used for further modelling of the experimental data. The quadratic model was selected in this study because it had the highest adjusted R² and predicted R² values.

Table 3 Adequacy of the model tested for the response

Source	Prob > F	Std. dev.	R^2	Adjusted R^2	Predicted R^2	Press	Remarks
Linear	<0.0001	0.092	0.8070	0.7749	0.7327	0.28	—
Interactive (2FI)	0.9998	0.110	0.8093	0.7033	0.5265	0.50	—
Quadratic	<0.0001	0.030	0.9880	0.9761	0.9311	0.07	Suggested
Cubic	0.2481	0.025	0.9965	0.9835	0.4904	0.54	Aliased

---

The final model obtained in coded variables in which X₁, X₂, X₃ and X₄ are in sequence representing the NMDG concentration, reaction temperature, reaction time and DG, is given as:

Y = 1.57 + 0.071X₁ + 0.041X₂ + 0.20X₃ + 0.15X₄ + 0.0025X₁X₄ + 0.015X₂X₃ + 0.017X₂X₄ − 0.0075X₃X₄ − 0.028X₁² − 0.043X₂² − 0.17X₃² − 0.056X₄² (5)

3.2 Statistical analysis

The ANOVA analysis not only shows the relationship between the independent variables and the dependent variables but also determines the significance of each term as indicated by their p-values. The results of this analysis showed that the model is highly significant at a probability level of p < 0.0001 (as shown in Table 4).

Table 4 ANOVA analysis of the quadratic model of the independent variables

Source	Sum of squares	DF	Mean square	F-value	Prob > F	Remarks
Model	1.04	14	0.074	82.61	<0.0001	Significant
X1	0.060	1	0.060	66.81	<0.0001
X2	0.020	1	0.020	22.20	0.0003
X3	0.49	1	0.49	546.03	<0.0001
X4	0.28	1	0.28	309.67	<0.0001
X1^2	5.207 × 10^−3	1	5.207 × 10^−3	5.78	0.0306
X2^2	0.012	1	0.012	13.52	0.0025
X3^2	0.18	1	0.18	203.95	<0.0001
X4^2	0.020	1	0.020	22.44	0.0003
X1 X2	0.000	1	0.000	0.000	1.0000
X1 X3	0.000	1	0.000	0.000	1.0000
X1 X4	2.500 × 10^−5	1	2.500 × 10^−5	0.028	0.8701
X2 X3	9.000 × 10^−4	1	9.000 × 10^−4	1.00	0.3346
X2 X4	1.225 × 10^−3	1	1.225 × 10^−3	1.36	0.2631
X3 X4	2.250 × 10^−4	1	2.250 × 10^−4	0.25	0.6251
Residual	0.013	14	9.012 × 10^−4
Lack of fit	0.013	10	1.262 × 10^−3
Pure error	0.000	4	0.000
Cor total	1.05	28

---

The F value for the model was found to be high (82.61) with a low probability value of p < 0.0001. This suggests that the computed Fisher’s variance ratio was significant and has a high degree of adequacy for the quadratic model. The value of R² was 0.9880 and this indicates that 95% of the experimental data were compatible. The adjusted R² value was 0.9761 and this confirms the significance of the model. The predicted R² value was 0.9311 suggesting a reasonably good correlation between the observed and predicted values. According to Mourabet et al.,²⁵ if the difference between the adjusted R² and predicted R² is within 0.20, the predicted R² is in reasonable agreement with the adjusted R². Moreover, the adequate precision is a measure of the signal to noise ratio and a ratio greater than a value of 4 is desirable.²⁵ The recorded adequate precision value in this study was 34, which indicates the presence of an adequate signal to noise ratio. The coefficient of variance (CV) value was found to be 2.07%, which is low and also indicates the accuracy and reliability of the experimental results.²⁶

The data were analysed to check the normality of the residuals and the normal probability plot of the studentised residuals is shown in Fig. 2. It can be observed that the data points fit a straight line reasonably. This suggests that the model is established and the results of the ANOVA analysis obtained in this study are adequate and valid.

Fig. 2 Plot of normal probability versus studentised residuals.

3.3 Analysis of the response surfaces and effects of independent variables

The interactions between the independent variables (NMDG concentration, reaction temperature, reaction time and DG) and the glucamine density in the functionalised adsorbent shown in Fig. 3(a–c) were optimised using the numerical optimisation technique. The density (mmol g⁻¹-adsorbent) was obtained with respect to the NMDG loaded onto the poly(VBC) grafted nylon-6 fibers. Fig. 3a shows a 3D response surface plot of the variation of the glucamine density with NMDG concentration and DG at a constant reaction time of 47 min and temperature of 81 °C. The glucamine density increased with the increase of the DG and NMDG concentrations until it achieved a maximum value at DG and NMDG concentrations of 121% and 11%, respectively. Further increases in the DG and NMDG concentrations beyond these values did not result in any increase of the glucamine density.

Fig. 3 Response surface plots of (a) NMDG concentration (wt%) versus DG (%); (b) NMDG concentration (wt%) versus temperature (°C); (c) NMDG concentration (wt%) versus time (min) on the glucamine density as the response.

The 3D response surface plot of the variation of the glucamine density with NMDG concentration and temperature at a constant reaction time of 47 min and DG of 121% is shown in Fig. 3b. The glucamine density increased with the increase of the NMDG concentration and temperature where both parameters affected the glucamine density in a similar manner.

A maximum glucamine density was recorded at the NMDG concentration of 11 wt% and temperature of 82 °C beyond which no further increase takes place despite the increase in parameter values.

Fig. 3c shows a 3D surface plot of the glucamine density versus NMDG concentration and time at the reaction temperature of 81 °C and DG of 121%. The increased NMDG concentration and time result in an increase of the glucamine density. The reaction time of 40 min gave a maximum glucamine density beyond which it starts to level off and reach saturation.

Fig. 4 shows the Ramp report, which summarises the optimum predicted values for the independent variables leading to the maximum glucamine density using the BBD model. The optimum reaction parameters for achieving a glucamine density of 1.72 mmol g⁻¹-adsorbent with a desirability function value of 1.000 are 10.6% NMDG concentration, 81 °C reaction temperature, 47 min reaction time and 121% DG.

Fig. 4 Ramp report showing the optimum parameters at 1.0 desirability level.

3.4 Experimental validation of the glucamine density

The predicted glucamine density value of 1.72 mmol g⁻¹-adsorbent was validated under the optimum reaction parameters in triplicate runs. The average experimentally determined glucamine density was found to be 1.70 mmol g⁻¹-adsorbent. This value agrees very well with the predicted value with a relatively minor deviation of 1.18%. Such agreement suggests that the optimisation of the glucamine containing poly(VBC) grafted nylon-6 using the BB design was successful.

3.5 Morphological properties

SEM images of the original nylon-6, poly(VBC) grafted nylon-6 and corresponding glucamine functionalised nylon-6 fibers are shown in Fig. 5. The diameter of the original nylon-6 fibers was increased from 15 μm (Fig. 5a) to 24 μm by grafting poly(VBC) (Fig. 5b) and to 30 μm (Fig. 5c) after introducing glucamine functional groups (1.7 mmol g⁻¹-adsorbent). The two-stage increase in the nylon-6 fiber diameter confirms the incorporation of poly(VBC) grafts and the addition of glucamine to the grafted benzyl group.

Fig. 5 SEM images of (a) original nylon-6 fibers, (b) poly(VBC) grafted nylon-6, 130% DG, and (c) glucamine functionalised poly(VBC) grafted nylon-6 fibers with 1.7 mmol g⁻¹-adsorbent.

3.6 Chemical structure

FT-IR spectra of the original, poly(VBC) grafted and glucamine functionalised nylon-6 fibers with various densities are given in Fig. 6 and 7. The characteristic peaks of the original nylon-6 are listed in Table 5. The intensities of the peaks for amide I at 1645 cm⁻¹, amide II at 1544 cm⁻¹ and other peaks belonging to nylon-6 were reduced after grafting with poly(VBC) and subsequent functionalisation with NMDG. This was followed by an increase in the poly(VBC) adsorption peak at 1264 cm⁻¹ representing the C–Cl stretching vibration of the chloromethyl group^27,28 and subsequent gradual reduction with the increase in the density of glucamine incorporated (Fig. 7). The decrease in the peak representing C–Cl is due to the substitution of –Cl by –NMDG groups, which also increased the broadness of the band in the range of 3350–3600 cm⁻¹ representing OH groups present in glucamine. The new vibration bands at 1072 and 1017 cm⁻¹ are characteristic to C–O, and C–N of NMDG and their intensity is a function of the glucamine density in the adsorbent. Hence, these results confirm that glucamine groups were successfully introduced to the benzyl groups of the poly(VBC) grafted nylon-6.

Fig. 6 FT-IR spectra of the original nylon-6, poly(VBC) grafted nylon-6 and NMDG functionalised poly(VBC) grafted nylon-6 fibers with different densities in the range of 1500–3700 cm⁻¹.

Fig. 7 FT-IR spectra of the original nylon-6, poly(VBC) grafted nylon-6 and NMDG functionalised poly(VBC) grafted nylon-6 fibers with different densities in the range of 600–1500 cm⁻¹.

Table 5 FT-IR bands for nylon-6 fibers

Band (cm^−1)	Assignment
3300	Hydrogen-bonded NH stretching
3086	NH Fermi resonance
2931	CH2 asymmetric stretching
2859	CH2 symmetric stretching
1645	Amide I
1544	Amide II
1369	Amide III + CH2 wagging
1264	Amide III + CH2 wagging
1236	CH2 wagging/twisting
1203	Amide III + CH2 wagging
1170	CONH skeletal motion
1121	CC stretching (amorphous)
1072	CC stretching
974	CONH in-plane (γ)
929	CONH in-plane (α)
728	Amide V (γ)
705	Amide V (α)

---

Similar glucamine characteristic bands were reported in the literature for the synthesis of polymer/clay nanocomposite ion exchange resins based on NMDG for arsenic removal.²⁹

3.7 Structural properties

Fig. 8 displays the diffraction patterns of the original, poly(VBC) grafted and glucamine functionalised poly(VBC) grafted nylon-6 fibers with various NMDG densities. The diffraction pattern of the original nylon-6 has the morphology of a semicrystalline polymer with a crystalline γ-form peak position at 2θ = 21.2°. This observation is in agreement with the literature.^30,31 The incorporation of poly(VBC) to nylon-6 caused a reduction in the crystalline peak intensity of the γ-form and the appearance of the α-form peaks at 2θ = 19.9° and 2θ = 23.5°, respectively. A further decrease in the peak intensity of the γ-form is clearly observed after functionalisation with NMDG until it is almost disappeared at a density of 1.7 mmol g⁻¹-adsorbent. This was accompanied by a marginal reduction in the peak intensities of the crystalline α-form.

Fig. 8 X-ray diffraction profiles of the nylon-6, poly(VBC) grafted nylon-6 and glucamine functionalised poly(VBC) grafted nylon-6 with various densities.

It is worth mentioning that the observation obtained from X-ray analysis is going along with the findings of FT-IR discussed earlier. Particularly, the incorporation of poly(VBC) onto nylon-6 fibers caused a major reduction in the overall crystallinity containing γ-crystalline and α-crystalline forms. This was followed by another decrease in the overall crystallinity caused by the loading of NMDG, which is dependent on the glucamine density of the final adsorbent. This was evident from the FT-IR spectra (Fig. 7) which showed an enhancement in the amorphous phase of the adsorption band at 1121 cm⁻¹ representing C–C. This trend is likely to be due to the decrease in the overall crystallinity of the glucamine containing fibers.

Furthermore, the bending mode of the CO–NH group at 974 cm⁻¹ corresponds to the characteristics of the crystalline γ-form and its intensity is also reduced with the increase in the density of NMDG in the poly(VBC) grafted nylon-6 fibers to the extent of disappearance at the density of 1.7 mmol g⁻¹-adsorbent. On the other hand, the peak intensity of the CO–NH bending mode at 929 cm⁻¹ associated with the α-crystalline form remains almost unchanged regardless of the density of glucamine in the adsorbent (Fig. 9). These observations confirm that the increase in the amount of glucamine groups loaded into poly(VBC) grafted nylon-6 fibers caused a transition in the crystalline structure from the γ-form to the α-form. The results suggest that the reduction in the overall crystallinity of the glucamine-functionalised fibers is due to the dilution of the crystalline structure with the amorphous NMDG-containing poly(VBC) grafts in the nylon-6 backbones.

Fig. 9 FT-IR spectra of the original nylon-6, poly(VBC) grafted nylon-6 and glucamine functionalised poly(VBC) grafted nylon-6 with different densities in the range of 810–910 cm⁻¹.

3.8 Thermal properties

Fig. 10 shows the DSC thermograms of the original nylon-6, poly(VBC) grafted nylon-6 and glucamine functionalised poly(VBC) grafted nylon-6 fibers and the obtained data are presented in Table 6. The first transition which appears below and close to 60 °C is likely to be due to a glass transition temperature (T_g) of nylon-6, which is masked by a moisture peak. The masked T_g peaks were shifted to higher values by the incorporation of the amorphous poly(VBC) grafts onto the nylon-6 fibers and the loading of glucamine onto the poly(VBC) grafted nylon-6 fibers. No changes in the masked T_g were caused by the increase in the density of glucamine in the adsorbent. All samples also displayed single endothermic melting peaks. The incorporation of poly(VBC) onto nylon-6 caused a substantial decrease in the melting temperature (T_m) of the original nylon-6 by 10.2 °C (Table 6). The introduction of glucamine to the poly(VBC) grafted nylon-6 showed a marginal reduction in the T_m, which increased again with higher glucamine density. The high reduction in the T_m after the grafting of VBC is likely caused by the major reduction in the overall crystallinity which accompanied the transformation of the crystal structure from the less stable γ-form to the more-stable α-form caused by the applied grafting procedure as previously discussed (section 3.7). This is also going along with the sharp decrease in the heat of fusion caused by the grafting which is equivalent to a 41.5% loss of the original crystallinity of the nylon-6 fibers. Interestingly, the change in the T_m was small in the functionalised adsorbent at a glucamine density of 1.7 mmol g⁻¹-adsorbent despite the transformation of most of the crystal structure in the γ-form to the α-form. This trend suggests that the main reason for the reduction in the T_m after functionalisation is the dilution effect caused by the incorporation of the amorphous glucamine-containing poly(VBC) onto nylon-6 fibers.

Fig. 10 DSC thermograms of the nylon-6, poly(VBC) grafted nylon-6 and glucamine functionalised poly(VBC) grafted nylon-6 fibers with various densities.

Table 6 Melting temperatures for the original nylon-6, poly(VBC) grafted nylon-6 and glucamine functionalised poly(VBC) grafted nylon-6 with various densities

Sample	Glucamine density (mmol g^−1 adsorbent)	Tm (°C)	ΔHf (J g^−1)
Nylon-6	0.0	217.9	60.1
VBC-grafted nylon-6	0.0	207.7	35.1
NMDG functionalised poly(VBC) grafted nylon-6 (1)	0.5	206.9	16.8
NMDG functionalised poly(VBC) grafted nylon-6 (2)	1.7	208.1	15.7

---

3.9 Thermal stability

Fig. 11 shows the TGA thermograms of the original nylon-6, poly(VBC) grafted nylon-6 and glucamine functionalised poly(VBC) grafted nylon-6 fibers. The TGA thermogram of the original nylon-6 remains stable up to 350 °C beyond which its thermal decomposition takes place at 445 °C. The grafted nylon-6 shows three stages of decomposition at temperatures of 315, 395 and 445 °C, respectively. These temperatures represent a 2-stage depolymerisation of poly(VBC) from the grafted fibers having two different crystalline (γ and α) forms and the thermal decomposition of the nylon-6 molecular chains, respectively. The glucamine-containing nylon-6 fibers show multiple decomposition stages. The first stage, which is obvious for the sample with the 1.7 glucamine density, starts in the temperature range of 80–200 °C and is due to the loss of water (dehydration) bound to the functional groups by H-bonding. This is followed three decomposition stages at 285, 385 and 445 °C which represent the decomposition of the glucamine group, depolymerisation of the poly(VBC) having a single crystalline α-form and decomposition of the molecular chains of nylon-6. The sample with a higher glucamine density seems to be less thermally stable than the one with a lower glucamine density. These observations suggest that the adsorbent remains thermally stable up to 250 °C. Moreover, it has sufficient thermal resistance for application in boron removal from solutions at high temperature.

Fig. 11 TGA thermograms of the nylon-6, poly(VBC) grafted nylon-6 fibers and glucamine functionalised poly(VBC) grafted nylon-6 fibers.

3.10 Mechanical properties

Fig. 12 shows the variation of tensile strength of the original nylon-6, poly(VBC) grafted nylon-6 and glucamine functionalised poly(VBC) grafted nylon-6 fibers with various densities. The tensile strength of the original nylon-6 decreased after the grafting of VBC as a result of the decrease in the crystallinity. However, the poly(VBC) grafted nylon-6 fibers recorded a high tensile strength after introducing NMDG at a density of 0.5 mmol g⁻¹-adsorbent beyond which it decreased gradually but at a slow pace. The rise in the tensile strength (at 0.5 mmol g⁻¹-adsorbent) is likely to be due to the formation of a network structure imparted by the glucamine to the surface of the fibers. On the contrary, the decrease in the tensile strength with further increase in the glucamine density is most likely caused by the deterioration of the mechanical properties of the grafted nylon-6 fibers by the prolonged immersion in 1,4-dioxane (solvent) to achieve the high glucamine density. For example, the grafted nylon-6 fibers were immersed in NMDG/1,4-dioxane solution for 5 minutes to achieve the 0.5 mmol g⁻¹-adorbent and 47 min to achieve the 1.7 mmol g⁻¹-adsorbent. A similar trend was observed for the variation of displacement with the increase in the glucamine density in the functionalised nylon-6 fibers. These results suggest that the newly prepared adsorbent retains a great part of the mechanical properties of the starting nylon-6 fibers despite the grafting with poly(VBC) and subsequent treatment with NMDG.

Fig. 12 The tensile strength and displacement of the original nylon-6, poly(VBC) grafted nylon-6 and glucamine functionalised poly(VBC) grafted nylon-6 fibers.

3.11 Adsorption test

Fig. 13 shows the adsorption capacity of boron onto the glucamine functionalised fibrous adsorbent as a function of pH. The highest adsorption capacity was achieved at pH values in the range of 5–9. A similar trend was also reported for the removal of boron from solutions onto NMDG functionalised magnetic micro-particles of silica³² and commercial resin, Diaion CRB 02.³³ The maximum adsorption capacity of 13.8 mg g⁻¹ was obtained at pH 7 and therefore the boron solution was adjusted to pH 7 in other experiments.

Fig. 13 Effect of pH on boron removal (initial boron concentration 100 mg L⁻¹, adsorbent dose 0.5 g, time 2 hours, temperature 30 °C and stirring speed 200 rpm).

Fig. 14 shows the variation of the removal efficiency of boron with the fibrous adsorbent dose at two different boron concentrations in boron solution. The removal efficiency increases with the increase in the adsorbent dose from 0.05 to 1.00 g until 100% removal was achieved at 0.5 g and 0.8 g for the treatment of boron solutions having concentrations of 100 mg L⁻¹ and 200 mg L⁻¹, respectively. This observation is due to the increase in the adsorption capacity with the increase in the surface area of the fibrous adsorbent. These results suggest that the new adsorbent is capable of complete removal of boron from solutions and the utilised adsorbent dose depends on the initial concentration of boron in solution. Moreover, it is anticipated that the glucamine-containing poly(VBC) grafted nylon-6 could be an alternative adsorbent for the effective removal of boron from solutions. The most important properties of the glucamine-containing poly(VBC) grafted nylon-6 fibrous adsorbent are presented in Table 7.

Fig. 14 Effect of adsorbent dosage on boron removal (initial solution concentration 100 and 200 mg L⁻¹, time 2 hours, temperature 30 °C, stirring speed 200 rpm, pH 7).

Table 7 Properties of the glucamine-containing poly(VBC) grafted nylon-6 fibrous adsorbent

Properties	Description
Matrix	Poly(VBC) grafted nylon-6
Physical form	Fibers
Av. fiber diameter	30 μm
Chelating group	NMDG
Glucamine density	1.70 mmol g^−1
Adsorption capacity	13.8 mg g^−1 adsorbent (pH = 7)

---

3.12 Stability of the prepared fibrous adsorbent

The fibrous adsorbent was eluted using HCl acid solution to release the bound borates and converted to the free base form again using a NaOH solution through several adsorption/desorption cycles. Fig. 15 shows the adsorption capacity versus the number adsorption/desorption cycles. As can be seen, the adsorption capacity of boron remained almost constant after 5 cycles. This indicates that the obtained glucamine-containing fibrous adsorbent has sufficient stability for multiple cycles of boron removal.

Fig. 15 Adsorption capacity versus the number of adsorption/desorption cycles (initial solution concentration 100 mg L⁻¹, adsorbent dose 0.5 g, temperature 30 °C, stirring speed 200 rpm and reaction time 2 h).

4. Conclusions

A new boron-selective adsorbent containing glucamine hosted by poly(VBC) grafted on nylon-6 obtained by radiation induced grafting of VBC on nylon-6 and subsequent treatment with NMDG was successfully prepared. The dependence of the glucamine density in the adsorbent on the functionalisation reaction parameters was analysed with RSM using Box–Behnken design. The adopted quadratic polynomial model enabled facile tuning of the content of the functional groups in the adsorbent. The RSM was found to be effective in predicting the yield of the functionalisation reaction and for optimization of the reaction parameters. The maximum level of the glucamine density was found to be 1.7 mmol g⁻¹-adsorbent and was obtained at the optimum parameters of 10.60% NMDG concentration, 81 °C reaction temperature, 47 min reaction time and 121% DG. The characterization of the prepared adsorbent using SEM, FT-IR, XRD, DSC and TGA analyses showed undoubtable evidence of the incorporation of glucamine groups in the poly(VBC) grafted nylon-6 fibers. The obtained fibrous adsorbent was found to retain reasonable mechanical (tensile strength = 125 MPa and displacement = 4.8 mm) and thermal properties (T_m = 208 °C) together with high thermal stability (250 °C). The preliminary performance of the fibrous adsorbent in boron adsorption showed complete boron adsorption (i.e. 100% removal efficiency) from aqueous solution was achieved using a dose of 0.5 g at an initial boron concentration of 100 mg L⁻¹, a reaction time of 2 h, temperature of 30 °C, stirring of 200 rpm and pH of 7. The adsorption capacity of boron remained almost constant after five cycles of adsorption/desorption. This suggests that the fibrous glucamine-containing adsorbent based on radiation grafted poly(VBC)/nylon-6 fibers is a promising adsorbent for application in boron removal from solutions.

Acknowledgements

T. M. Ting wishes to acknowledge sponsorship of the research by MOSTI E-Science Fund (03-03-01-SF0058) and HLP, JPA. M. M Nasef is also grateful to the financial support from UTM research fund (grant no. 2543.05H16).

References

1. N. Kabay, I. Yilmaz, S. Yamac, M. Yuksel, U. Yuksel, N. Yildirim, O. Aydogdu, T. Iwanaga and K. Hirowatari, Desalination, 2004, 167, 427–438.
2. N. Hilal, G. J. Kim and C. Somerfield, Desalination, 2011, 273, 23–35.
3. M. M. Nasef, M. Nallappan and Z. Ujang, React. Funct. Polym., 2014, 85, 54–68.
4. N. Kabay, S. Sarp, M. Yuksel, Ö. Arar and M. Bryjak, React. Funct. Polym., 2007, 67, 1643–1650.
5. N. Bicak, N. Bulutcu, B. F. Senkal and M. Gazi, React. Funct. Polym., 2001, 47, 175–184.
6. Y. Inukai, Y. Tanaka, T. Matsuda, N. Mihara, K. Yamada, N. Nambu, O. Itoh, T. Doi, Y. Kaida and S. Yasuda, Anal. Chim. Acta, 2004, 511, 261–265.
7. H. Liu, X. Ye, Q. Li, T. Kim, B. Qing, M. Guo, F. Ge, Z. Wu and K. Lee, Colloids Surf., A, 2009, 341, 118–126.
8. S. Nishihama, Y. Sumiyoshi, T. Ookubo and K. Yoshizuka, Desalination, 2013, 310, 81–86.
9. H. Parschová, E. Mištová, Z. Matějka, L. Jelínek, N. Kabay and P. Kauppinen, React. Funct. Polym., 2007, 67, 1622–1627.
10. M. Palencia, M. Vera and E. Combatt, J. Appl. Polym. Sci., 2014, 131 DOI:10.1002/app.40653.
11. J. Sanchez, B. L. Rivas, E. Nazar, M. Bryjak and N. Kabay, J. Appl. Polym. Sci., 2013, 129, 1541–1545.
12. S. Samatya, N. Kabay and A. Tuncel, J. Appl. Polym. Sci., 2012, 126, 1475–1483.
13. M. M. Nasef and O. Güven, Prog. Polym. Sci., 2012, 37, 1597–1656.
14. M. R. Awual, S. Urata, A. Jyo, M. Tamada and A. Katakai, Water Res., 2008, 42, 689–696.
15. Y. Hamabe, R. Matsuura, A. Jyo, M. Tamada and A. Katakai, React. Funct. Polym., 2009, 69, 1–8.
16. M. M. Nasef and E.-S. A. Hegazy, Prog. Polym. Sci., 2004, 29, 499–561.
17. K. Ikeda, D. Umeno, K. Saito, F. Koide, E. Miyata and T. Sugo, Ind. Eng. Chem. Res., 2011, 50, 5727–5732.
18. H. Hoshina, N. Seko, Y. Ueki and M. Tamada, J. Ion Exch., 2007, 18, 236–239.
19. T. M. Ting, H. Hoshina, N. Seko and M. Tamada, Desalin. Water Treat., 2013, 51, 2602–2608.
20. N. K. Pramanik, R. S. Haldar, Y. K. Bhardwaj, S. Sabharwal, U. K. Niyogi and R. K. Khandal, Radiat. Phys. Chem., 2009, 78, 199–205.
21. M. C. Gupta and R. R. Pandey, J. Polym. Sci., Part A-1: Polym. Chem., 1988, 26, 491–502.
22. G. E. P. Box and K. B. Wilson, J. Roy. Stat. Soc. B, 1951, 13, 1–45.
23. D. C. Montgomery and G. C. Runger, Applied Statistics and Probability for Engineers, 5th edn, John Wiley and Sons, Inc., USA, 2011.
24. G. E. P. Box and D. W. Behnken, Technometrics, 1960, 2, 455–475.
25. M. Mourabet, A. El Rhilassi, H. El Boujaady, M. Bennani-Ziatni, R. El Hamri and A. Taitai, Appl. Surf. Sci., 2012, 258, 4402–4410.
26. J. Prakash Maran, S. Manikandan and V. Mekala, Ind. Crops Prod., 2013, 49, 304–311.
27. W. F. Wang, T. L. Tan, B. L. Tan and P. P. Ong, J. Mol. Spectrosc., 1996, 175, 363–369.
28. E. Yaacoub and P. L. Perchec, React. Polym., 1988, 8, 285–296.
29. B. F. Urbano, B. L. Rivas, F. Martinez and S. D. Alexandratos, React. Funct. Polym., 2012, 72, 642–649.
30. H. R. Pant, W.-.i. Baek, K.-T. Nam, I.-S. Jeong, N. A. M. Barakat and H. Y. Kim, Polymer, 2011, 52, 4851–4856.
31. C. B. Giller, D. B. Chase, J. F. Rabolt and C. M. Snively, Polymer, 2010, 51, 4225–4230.
32. B. Tural, Clean - Soil, Air, Water, 2010, 38, 321–327.
33. İ. Kıpçak and M. Özdemir, Environ. Technol., 2010, 31, 327–335.

---