Defluoridation technology for drinking water and tea by green synthesized Fe3O4/Al2O3 nanoparticles coated polyurethane foams for rural communities

Fluoride (F) contaminated ground water poses a serious public health concern to rural population with unaffordable purification technologies. Therefore, development of a cost-effective, portable, environment and user-friendly defluoridation technique is imperative. In the present study, we report on the development of a green and cost-effective method that utilizes Fe3O4 and Al2O3 nanoparticles (NPs) that were synthesized using jojoba defatted meal. These NPs were impregnated on to polyurethane foam (PUF) and made into tea infusion bags. The Al2O3 NPs-PUF displayed a higher water defluoridation capacity of 43.47 mg g−1 of F as compared to 34.48 mg g−1 of F with Fe3O4 NPs-PUF. The synthesized Al2O3-PUF infusion bags removed the F that was under the permissible limit of 1.5 mg L−1. The sorption experiments were conducted to verify the effect of different parameters such as pH, contact time, size of PUF and initial F concentration. The different properties of adsorbent were characterized using a combination of FESEM, EDX, XRD and FTIR techniques, respectively. The calculated total cost per NPs-PUF pouch developed is as low as US $0.05, which makes the technology most suitable for rural communities. This paper will be beneficial for researchers working toward further improvement in water purification technologies.


Results and Discussion
Morphology and Chemical composition. The Fe 3 O 4 and Al 2 O 3 NPs were synthesized from Simmondsia chinensis (jojoba) defatted meal extract by a green synthesis route as described in experimental methods. The surface morphology and elemental composition of synthesized Fe 3 O 4 and Al 2 O 3 NPs was characterized using FESEM and EDX spectrum analysis. Figure 1 (Fig. 1a). The surface morphology of Al 2 O 3 NPs appeared to be flakes in nature with irregular shape (Fig. 1c). The composition of NPs was further analyzed by EDX elemental mapping. The elemental composition from EDX analysis confirmed that the Fe 3 O 4 NPs sample has O (36.99%), Fe (54.34%) and Si (8.66%) (Fig. 1b). Likewise, the EDX measurements with Al 2 O 3 NPs showed the presence of O (53.11%), Al (25.09%) and Si (21.80%) (Fig. 1d). Based on the EDX spectrum analysis, it was confirmed the presence of both Fe and Al elements in the samples. Figure 1(e and g) shows the FESEM images of PUFs after impregnation with Fe 3 O 4 and Al 2 O 3 NPs at different magnifications, indicating the binding of NPs. It is clear from the FESEM images that PUF has closed cell structure with NPs coated on its wall surfaces. The F peak in EDX spectrum showed the F adsorption process through Fe and Al NPs-PUF, confirming the F adsorption by an adsorbent (Fig. 1f and h).
The tea bag covering filter paper surface morphology was studied by FESEM before and after F adsorption and is presented in Fig. 2(a and c). The FESEM analysis revealed that the Al-NPs were uniformly coated onto tea bag filter paper. The EDX spectrum confirmed the presence of Al element in Al-PUF tea bag sample before and after adsorption ( Fig. 2b and d). The presence of F ions peak along with Al peak confirmed the F adsorption from tea infusions (Fig. 2d).
Phase composition. Phase purity and crystallinity of the prepared Fe 3 O 4 and Al 2 O 3 NPs was recognized through XRD analysis. The XRD patterns of synthesized Fe 3 O 4 and Al 2 O 3 NPs are shown in Fig. 2(e and f). The three major diffraction peaks of synthesized Fe 3 O 4 NPs were detected at 2θ = 35.51°, 62.59° and 30.07° (Fig. 2e), which are assigned to the crystal planes of (311), (440) and (220), respectively. The achieved peaks were similar to the standard patterns of JCPDS file no: 00-019-0629, which stated the crystallographic system of cubic structure of Fe 3 O 4 . For the Al 2 O 3 NPs, intense diffraction peaks at 66.80°, 45.62° and 36.94° were observed, which corresponded to the planes (240), (−422) and (221), respectively (Fig. 2f). The XRD patterns declared the monoclinic crystal phase of Al 2 O 3 NPs (00-011-0517). The average crystalline size of both NPs can be determined using the Debye-Scherrer equation 25 . The Debye-Scherrer equation: where D is the crystallite size, k is Scherrer constant (0.9), λ is the X-ray wavelength of radiation for Cu Kα (0.154 nm), β hkl is the full-width at half maximum (FWHM) and θ hkl is the diffraction angle.  Fig. 3a. Characteristic peaks were observed at 556 and 3670 cm −1 in the spectra for Fe 3 O 4 which are assigned to the stretching of metal-oxygen because of the Fe-O and O-H groups, respectively (Fig. 3a) 31 . The peaks occurred at 625, 1040, and 1120 in the PUF are assigned to the C-H, C-O-C stretch of ester and C-O stretch, respectively. The sharp peaks observed at 1640, 1730, 2840, 3300, and 3636 cm −1 in the PUF sample showed the presence of N-H stretch of urea, C=O stretch of urethane, C-H stretch, N-H stretch of urethane and urea, and O-H stretch, respectively (Fig. 3a). All the urethane functional group peaks were observed in PUF 32 . All the characteristic peaks observed in the Fe 3 O 4 NPs and PUF were also observed in Fe 3 O 4 NPs-PUF samples. The increased intensity of O-H and N-H band in Fe 3 O 4 NPs-PUF spectrum before F adsorption indicated Fe chelation of N-H groups in PUF. The decreased intensity of Fe-O band after F adsorption can be associated with interaction with the F ions, the similar behavior of Fe-O was also reported in literature 33 (Fig. 3a). The decrease in the intensity of O-H bond after F adsorption indicated the replacement of hydroxyl ions by the F ions 34 . The presence of Fe-O band in the spectra of Fe 3 O 4 NPs-PUF before and after F adsorption confirmed that Fe 3 O 4 NPs was complexed by PUF. The fact that no significant changes were observed in Fe 3 O 4 NPs-PUF spectra before and after F adsorption showed that no significant structural changes occurred in the Fe 3 O 4 NPs-PUF sample during the adsorption process.

Adsorption of Fluoride.
Effect of pH. The pH of the aqueous solution plays a noteworthy role in the F removal during both dip adsorption and batch studies. The surface charge of the mineral oxides is positive when pH value is below pH zero point charges (ZPC) and negatively charged when pH value is above the ZPC. The F adsorption by Fe 3 O 4 NPs-PUF and Al 2 O 3 NPs-PUF was observed to be strongly pH dependant. The percentage F removal increased with increasing pH up to 5 and 6 for Fe 3 O 4 NPs-PUF and Al 2 O 3 NPs-PUF, respectively. But F removal percentage is decreased in the pH range of 5.0-9.0 for Fe 3 O 4 NPs-PUF and 6.0-9.0 for Al 2 O 3 NPs-PUF. These results demonstrated the reduction in F removal upon enhancing the pH above 5 and 6 for Fe 3 O 4 and Al 2 O 3 NPs-PUF, respectively (Fig. 4a). In acidic pH conditions, the formation of hydrofluoric acid (HF) is responsible for the reduction of F adsorption. Under alkaline conditions, F removal declined because of the competition between F ions and hydroxyl ions for the active surface sites. In addition, the electrical repulsion among negatively charged adsorbent surface sites is probably responsible for the low absorbance.
Initial fluoride concentration and contact time effect. The adsorption of F ions reduced as the initial F concentration increased (Fig. 4b). The percent fluoride removal was found to be 93 and 96.3% for Fe 3 O 4 NPs and Al 2 O 3 NPs-PUF, respectively from the initial 2 mg L −1 F concentration, which further decreased to 20.9 and 25.2% for Fe 3 O 4 NPs and Al 2 O 3 NPs-PUF, respectively from the initial 10 mg L −1 F concentration at a contact time of 80 min (Fig. 4b). The variation in the percent F removal may be due to the decline in the number of available adsorption sites as they saturated at a excess F concentration. Adsorption behavior was studied as a function of contact time from 20 to 100 min with Al 2 O 3 NPs-PUF size 6 × 6 cm −1 at pH 6 at 30 °C (Fig. 4c). It clear from the above results that the adsorption enhances with time and an equilibrium state is attained after a contact time of 80 min.
Effect of the varying PUF size on F adsorption. The effect of varying Al 2 O 3 NPs-PUF sizes on F adsorption was evaluated using 2 mg L −1 initial F concentration and pH 6 ( Fig. 4c). With the increase in the size of Al 2 O 3 NPs-PUF from 2 to 6 cm 2 , the percent F removal also increased from 58.50 to 96.50%. The percent F removal by Fe 3 O 4 NPs-PUF also increased up to 93% from initial 2 mg L −1 F concentration at pH 5. The presence of extra NPs on surface with increase in PUF size allowed efficient interaction resulting in enhanced interaction and overall percent F removal. Fluoride concentration in tea infusions. The F levels are substantially found in all black, green and jasmine tea samples tested (Table 1). All tea products tested exceeded the permissible limit of 1.5 mg L −1 of F. F concentrations in leaf tea were considerably more than in bagged tea drinks. The F concentration in black tea was detected to be more than green and jasmine tea samples.
Fluoride removal from tea infusions. The defluoridation studies were carried out by simply dipping the Al 2 O 3 -PUF and Fe 3 O 4 -PUF infusion bags in 100 ml of tea samples, respectively. The measured F concentration after the defluoridation process using Al 2 O 3 -PUF infusion bags was found to be under the permissible limit (1.5 mg L −1 ). However, use of Fe 3 O 4 -PUF infusion bags did show defluoridation but failed to show permissible F levels in all tea infusions tested (Table 1).
Adsorption kinetics. The adsorption efficiency is illustrated using a variety of kinetic models. The adsorption kinetics was studied with pseudo-first-order and pseudo-second-order models. The data obtained was applied to pseudo-first-order and pseudo-second-order models to explain the adsorption kinetics of F ions on the Fe 3 O 4 and Al 2 O 3 NPs-PUF. The pseudo-first-order kinetic model is expressed by following eq. (2) 37 .  where q t and q e signify the quantities of F adsorbed (mg g −1 ) at time t and at equilibrium, respectively and k 1 (h −1 ) is the first-order reaction rate constant. The pseudo-second-order reaction is expressed by following eq. (3) 37 where k 2 (mg g −1 h −1 ) is the pseudo-second-order rate constant for F adsorption. The slope and intercept for both kinetic models were obtained by the linear kinetic plots, and kinetic parameters were determined as shown in Table 2. The obtained data demonstrated that the pseudo-second-order model fitted better for the adsorption study with highest correlation coefficient values (R 2 = 0.996 and 0.997 for Al 2 O 3 and Fe 3 O 4 NPs-PUF) than pseudo-first-order model. Both the Al 2 O 3 and Fe 3 O 4 NPs-PUF materials followed pseudo-second-order kinetics reveling that the F ions uptake takes place by means of chemisorption processes.

Adsorption isotherm studies.
To quantify the defluoridation capacity of NPs-PUF, three important isotherms were adopted. The experimental data obtained for the F concentration (2 mg L −1 ) at constant temperature and pH 6 and 5 for Al 2 O 3 and Fe 3 O 4 NPs-PUF were fitted to three commonly used isotherm models, such as Langmuir, Freundlich, and Temkin.
The Langmuir isotherm describes the monolayer adsorption and is shown in the linear form the following eq. (4) 37 : e e e 0 0 where C e is the equilibrium concentration of adsorbate (mg L −1 ), q e is the amount of F adsorbed at equilibrium (mg g −1 ), Q° is the adsorption for a complete monolayer (mg g −1 ), and b is the Langmuir isotherm constant (L mg −1 ). Figure 5a shows that experimental data fitted well with the Langmuir isotherm, maximum adsorption capacity was found to be 43.47 and 34. where k F is the Freundlich isotherm constant (mg g −1 ) and n is the adsorption intensity. Figure 5b shows the linear plots of Freundlich isotherm of F ions adsorbed on the Al 2 O 3 and Fe 3 O 4 NPs-PUF. The values of n > 1 represent the favorable adsorption condition and the calculated n value in the present study was calculated to be 1.78 and 1.35 for Al 2 O 3 and Fe 3 O 4 NPs-PUF, respectively that proves the favorable isotherm. The Temkin isotherm demonstrates as adsorbent-adsorbate interaction. A linear plot between q e and log C e demonstrates the Temkin isotherm as shown in Fig. 5c, which is defined by the following eq. (6) 38 with hydroxyl ions (Fig. 3). This may be due to the similar ionic radius of the iso-electronic OH and F ions 39 . NPs impregnation and F adsorption mechanism on the PUF is illustrated as shown in Fig. 6. However, no further structural changes were observed in NPs-PUF samples after F adsorption that indicates the mechanism of F adsorption occurred through ion exchange process, i.e., OH were replaced by F ions in the adsorption process. Cost of NP-PUF bag. The materials required for the development of NPs-PUF pouches were easily available and inexpensive PUF, metal salt and empty tea bags. The estimated cost of each NPs-PUF is estimated to be US $0.05 (Table 3). This proves that the cost of a developed defluoridation technology is most affordable to the rural population and the areas in the resource limited settings.

Conclusion
In this study, we reported on an inexpensive defluoridation technique that utilizes Fe 3 O 4 and Al 2 O 3 NPs that were green synthesized using jojoba defatted meal extract as reducing agent. These NPs were impregnated in PUF and fabricated tea infusion bags that were highly efficient in defluoridation of water and tea samples. The FTIR studies revealed that ion-exchange mechanism takes place between hydroxyl ions of NPs-PUF and F ions in samples. High F levels in tea infusion bags with black, green and jasmine tea were defluorinated to permissible F limits using Al 2 O 3 NPs-PUF tea-bag like pouches. The developed technique reported in this study has the advantages of high F removal capacity, ease of operation, portability, portability, environmental friendliness and low cost and thus making this approach most desirable to resource limited settings, especially in the rural areas with high ground water F contaminations. We believe that present study provides an affordable solution for F removal for rural and poor population for health and safety.

Methods
Chemicals and materials. All chemicals and reagents used in this study were of analytical grade. Plant material was collected from AJORP (Association of Rajasthan for Jojoba Plantation and Research Project), Jaipur, Rajasthan (India). Ferrous sulfate (FeSO 4 ) and aluminum nitrate (Al(NO 3 ) 3 ) precursor were obtained from Himedia, India. Sodium fluoride (NaF) was also supplied by Himedia, India and F stock solution (100 mg L −1 ) was prepared by adding NaF (0.0221 g) to millipore water (100 ml). Three common tea varieties, such as black, green and jasmine tea were procured from the local supermarket (Rajasthan, India). PUF and tea bag filter paper were procured from local suppliers and utilized after cleaning with millipore water. All experiments were carried out using Millipore ultrapure water.   Preparation of Simmondsia chinensis (jojoba) defatted meal extract. Jojoba was selected for green synthesis of Fe and Al NPs because of its abundant cost-effective and easy availability as a waste byproduct of oil extraction process. Green synthesis of NPs was performed as previously reported, with slight variations 41 . Prior to NPs synthesis, defatted jojoba seed meal was obtained. For this, the seed's were oven dried at 60 °C for 1 h and ground in a grinder. The resulting seeds powder was then refluxed in a soxhlet extractor for 24 h with in petroleum ether (1:6 w/v) for extracting oil. After oil extraction, the residual powder was termed as defatted jojoba meal and dried at room temperature for further use. Next, 10 g of defatted seed meal was added into 100 ml deionized water and boiled at 80 °C for 25 min. After cooling, the suspension obtained was filtered using Whatman's No.1 filter paper and stored at 4 °C. The filtrate was further utilized as reducing and stabilizing agent for NPs synthesis. Complete reduction of FeSO 4 to Fe + ions was confirmed by the color transformation from brown to black. The suspension was then centrifuged at 10,000 rpm for 15 min and the pellet obtained was repeatedly washed with millipore water and oven dried at 100 °C. For Al 2 O 3 NPs synthesis, aluminum nitrate (Al(NO 3 ) 3 ) was added into seed meal extract with 1:3 ratio (w/w) and allowed constant stirring at room temperature. The mixture obtained was microwave heated at 540 W for 7 min, which yield a yellow brown precipitate that was later centrifuged. The precipitate was washed with millipore water followed by methanol and dried at 100 °C in oven.

Preparation of Fe 3 O 4 and Al 2 O 3 NPs -PUF. The impregnation of Fe 3 O 4 and Al 2 O 3 NPs onto the PUF was
performed through dip adsorption method 42 . The NPs were suspended in 100 ml of distilled water for sonication. Then PUF were cut into a size of 2 × 2, 4 × 4 and 6 × 6 cm 2 with 3 mm thickness for impregnation process. For impregnation, 2 × 2 cm 2 size PUF was placed in 0.1 g NPs solution subjected to constant stirring at 200 rpm for 24 h at 30 °C in a shaker. Finally, the resulting NPs-PUF product was repeatedly washed with distilled water twice to remove un-anchored NPs on PUF. Thus obtained product (NPs-PUF) was dried at 80 °C in an oven. By increasing PUF sizes to 4 × 4 cm 2 , the NPs concentration was also appropriately increased to 0.2 g NPs concentration and thus maintained the NPs to PUF area ratio.
Adsorption experiments. The F adsorption experiments were carried out using a series of F concentrations, such as 2, 4, 6, 8 and 10 mg L −1 with 50 ml solution and with different NPs-PUF sizes in flasks. The contact time was varied to 20, 40, 60, 80 and 100 min and the flasks were placed in shaker at 140 rpm. Effect of pH on the F adsorption was studied in pH range of 2-9. At the end of adsorption process, the residual F concentration was determined by fluoride ion meter (Thermo Scientific Orion, USA). The removal efficiency of the adsorbent was calculated using following eq. (7) 37 : Removal efficiency (%) C C /C 100 where C o and C e (mg L −1 ) are the initial and equilibrium concentrations of F. The post adsorption NPs-PUF were removed and dried in a oven for further characterization using FESEM, EDX, and FTIR techniques.
NPs-PUF tea bag. For F removal from tea, easy to use tea-bag like pouches containing Al 2 O 3 NPs impregnated 6 × 6 cm 2 PUF and tea bag filter paper were prepared. The tea bag filter paper was impregnated through dipping in NPs solution overnight at 30 °C. For F adsorption, the Al-PUF bag was dipped in 100 ml of prepared tea for 5 min and further analyzed for F removal. Similar tea-bag like pouches were also prepared using Fe 3 O 4 NPs impregnated PUFs and the tea bag filter paper, further utilized for F remediation.
Preparation of tea infusions. Different tea bags, tea infusions and novel Al-PUF tea bag designed for F removal are shown in Fig. 7. The tap water was previously analyzed for F concentrations that were taken into consideration during the experiments. Each tea infusion was brewed for 2.5 min in 100 ml water at 95 to 98 °C, as the usual tea making time reported is 2 to 3 min 12 . After 2.5 min, tea infusions were filtered and allowed to cool for the F analysis.
Replication of the experiment. Each F adsorption experiment was conducted thrice and all the data share the average values of triplicate experiments.
Characterization of adsorbent. The surface structure of Fe 3 O 4 and Al 2 O 3 NPs were observed by FESEM (MIRA3 TESCAN). Morphology of tea bag filter paper, pure PUF and NPs-PUF were characterized by FESEM. The elemental composition of both NPs was identified by EDX analysis. Also, the composition of NPs impregnated tea bag filter paper and NPs-PUF before and after F removal (post-adsorption) was analyzed. Phase identification and crystal structures of the NPs were characterized using an X-ray diffractometer (XRD Bruker D8 Discover). The surface charge of Fe 3 O 4 and Al 2 O 3 NPs was characterized by the zeta (ζ) potential. The isoelectric point of NPs was identified by titrating the ζ-potential over the pH range of 2-9. The pH of the solutions was adjusted by adding H 2 SO 4 or NaOH. FTIR spectroscopy was carried out for Fe 3 O 4 and Al 2 O 3 NPs, NPs-PUF before and after F adsorption process for proposing an F adsorption mechanism. The thermal stability of the material was determined by Thermogravimetric analysis (TGA). The effect of NPs on the thermal properties of prepared NPs-PUF was analyzed.
Data availability. No datasets were generated or analyzed during the current study.