Hydrous TiO₂@polypyrrole hybrid nanocomposite as an efficient selective scavenger for the defluoridation of drinking water

Abstract

An adsorptive process for the defluoridation of drinking water was performed using a hybrid nanocomposite of hydrous titanium oxide@polypyrrole (HTiO₂@PPy), as a scavenger. The adsorbent was successfully fabricated via facile in situ chemical oxidative polymerization of pyrrole monomer in aqueous media in which HTiO₂ nanoparticles were suspended. The developed adsorbent was characterized using various spectro-analytical techniques viz. BET, FTIR, FE-SEM, STEM, EDX, TGA and ZETA SIZER. Relatively high BET surface area (98.17 m² g⁻¹) and pH_pzc (∼8.4) values were obtained for HTiO₂@PPy. The synergistic effect of both the counterparts (PPy and HTiO₂) of the nanocomposite rapidly enhanced the F⁻ adsorption process. A noteworthy rapid fluoride uptake best described by the pseudo-second-order kinetic model was observed (equilibrium attainment within 5–30 min). The Langmuir model best described the isotherm data with a maximum adsorption capacity of 31.93 mg g⁻¹ at 25 °C and pH 6.5 (±0.2). Thermodynamic and activation parameters provided evidence of the spontaneous, endothermic and physical nature of the adsorption process. The selectivity of HTiO₂@PPy for F⁻ sorption was significant in the presence of Cl⁻, NO₃⁻, HCO₃⁻, SO₄²⁻ and PO₄³⁻ co-existing ions and noteworthy reusability for up to three regeneration cycles was achieved. Electrostatic interactions and ion-exchange were proposed to be the possible underlying mechanisms for the adsorption of F⁻ by HTiO₂@PPy nanocomposite. Thus, HTiO₂@PPy is anticipated to serve as an efficient scavenger for the defluoridation of drinking water.

---

1. Introduction

Fluoride (F⁻) is a very well-known hazardous anion found in ground and surface waters in various forms as a result of climatic and anthropogenic activities.¹ The high concentration of F⁻ ions in groundwater is primarily contributed by the natural dissolution of fluoride-rich minerals, along with the industrial wastewater discharged from mining, plating and semiconductor factories, which consequently results in a more severe water pollution especially in African, Middle Eastern and Asian countries where groundwater is still a primary source of drinking water.^2–4 Nevertheless, the fluoride toxicity and its detrimental effects on muscle, brain, lung, kidney, thyroid, reproduction and enzymes along with dental and skeletal fluorosis are well reported.¹ Therefore, it has become globally important to bring down the F⁻ levels in the water below its maximum permissible limit (1.5 mg L⁻¹) which has been set by the World Health Organization (WHO).⁵

Over the past few decades, several techniques including precipitation–coagulation, adsorption, ion exchange, reverse osmosis and electrodialysis have been applied to remove F⁻ ions from groundwater.^6–8 Although each technique has its own inherent merits and demerits, adsorption is still considered as one of the most efficient, environmentally benign, economical and robust methods due to its relatively low cost, high efficiency, flexibility, easy handling and simple design.⁷ Several studies have focused on various engineered adsorbents for F⁻ removal from aqueous solutions, including activated and impregnated alumina, activated carbon, carbon nanotubes, clays, minerals, metal/mixed metal oxides, rare earth oxides, polymeric materials and resins.⁹ However, to date, some drawbacks associated with their applications such as a narrow working pH range, low selectivity and poor mechanical strength are still unresolved.^9,10 An ideal F⁻ adsorbent must have some preferred characteristics, such as low cost, good adsorption capacity, rapid uptake, easy regeneration and superior physical characteristics such as an inability to clog the pores of the filter.⁷ Unfortunately, most current inorganic adsorbents seldom have all of these features. Although nanomaterials have high surface areas, yet they lack various adsorbing functional groups. On the contrary, organic polymers hold a large number of polyfunctional groups but still their small specific area due to aggregation, limits their adsorptive potential.^11,12 Hence, the synthesis of hybrid/core–shell nanocomposites as adsorbents with both polyfunctional groups and the high surface area has recently gained considerable attention in wastewater treatment.¹³

Titanium-derived adsorbents are well known to have a good defluoridation potential.^13–16 However, the majority of these titanium-derived adsorbents were either prepared by high-temperature hydrolysis or by doping with expensive rare earth metal ions. Therefore, in the present work we used a simple and economic precipitation technique to synthesize hydrous titanium oxide nanoparticles (HTiO₂ NPs) using TiCl₄ as a precursor and further improvised its F⁻ adsorptive potential using a polymer.

Polypyrrole (PPy) is a well known conducting polymer that has fascinated the researchers for various applications due to its facile synthesis, high electrical conductivity, excellent biocompatibility, ion exchange and redox properties.¹⁷ PPy chains have positively charged nitrogen atoms that may easily become major active sites for the adsorption of anions like F⁻ and also exhibits an anion-exchanger behavior due to the presence of exchangeable counter anions doped in its chains. However, the major drawback associated with PPy is its poor dispersion in water and tendency to aggregate into irregular morphology due to strong π–π* interactions between the chains.¹⁸ Several researchers have studied fluoride removal properties of PPy and its composites. Karthikeyan et al. synthesized PPy, PPy/chitosan and PPy/alumina composites and obtained F⁻ removal capacities of 6.37, 6.67 and 8 mg g⁻¹, respectively.^17,19,20 Later, Bhaumik et al. also synthesized a PPy/Fe₃O₄ magnetic nanocomposite and obtained a F⁻ adsorption capacity of 17.6–22.3 mg g⁻¹.¹⁸ Moreover, quite recently, in our previous work, we had synthesized a PPy/HSnO nanocomposite by incorporating hydrous tin oxide and despite of the fact that hydrous tin oxide (HSnO) does not usually show great fluoride removing properties on its own, we successfully obtained a reasonably good F⁻ removal capacity of 26.16 mg g⁻¹.²¹ However, the results obtained in our previous work, enforced us to further explore and improvise the fluoride removing properties of PPy based nanocomposites by incorporating a new hydrous metal oxide which already has shown good defluoridation potential in literature (HTiO₂ NPs). Hence, in our present work it was anticipated that PPy could impart more surface active functional groups and hence can be used to tailor the adsorptive properties of the synthesized HTiO₂ NPs and simultaneously π–π* aggregation of PPy chains can also be prevented by incorporation of HTiO₂ NPs through organic–inorganic hybrid/core–shell approach.¹⁸

Hence, in the present work, a polypyrrole enwrapped hydrous titanium oxide core–shell hybrid (HTiO₂@PPy) nanocomposite was synthesized via an in situ chemical oxidative polymerization approach and characterized using various microscopic and spectroscopic characterization techniques. The effects of parameter such as adsorbent dose, solution pH, contact time, initial concentration, coexisting anions and temperature on the F⁻ removal efficiency of HTiO₂@PPy were investigated. Adsorption kinetics and isotherm models were established along with the evaluation of activation energy and thermodynamic parameters. The underlying mechanism of F⁻ removal by HTiO₂@PPy was evaluated with the help of various spectro-analytical techniques. Lastly, its regenerability, reusability, and real field applicability were tested.

2. Experimental procedures

2.1. Reagents

Pyrrole monomer (Py), obtained from Sigma-Aldrich, Germany was purified by vacuum distillation before use. Titanium(IV)tetrachloride (TiCl₄), anhydrous iron(III)chloride (FeCl₃), sodium fluoride (NaF), sodium hydroxide, hydrochloric acid, ammonia and all other chemicals were used as received from Sigma-Aldrich, Germany. Ultrapure deionized water (resistivity > 18.5 MΩ cm) was used for all the experiments. A F⁻ stock solution (1000 mg L⁻¹) was prepared by dissolving 2.210 g of NaF in 1000 mL of deionized water.

2.2. Adsorbent synthesis

Synthesis of HTiO₂ NPs. HTiO₂ NPs were synthesized using the method reported by Debnath et al.²² In brief, 30.0 mL of liquid TiCl₄ was injected very slowly in 1.0 L of deionized water under a fume hood with constant mechanical agitation at 1000 rpm. The white colored solution that appeared was treated with (2 : 8) dilute NH₃ solution till the pH of the supernatant liquid reached ∼6.5. The white colored precipitate obtained was allowed to settle for 1 h, filtered and then thoroughly washed with deionized water. The precipitate obtained was dried in the oven for 48 h at 70 °C and pulverized to obtain homogeneous HTiO₂ powder.

Synthesis of HTiO₂@PPy. HTiO₂@PPy was synthesized via an in situ chemical oxidative polymerization technique using FeCl₃ as an oxidant. Typically, 0.2 g of HTiO₂ NPs were dispersed in 50 mL of deionized water and stirred for 1 h for better dispersion. Then 0.2 mL of Py was syringed into it and again stirred for 30 min. 1.35 g of FeCl₃ was then added into the above solution; the reaction mixture was stirred for 4 h and then left unstirred for polymerization to proceed overnight. The obtained HTiO₂@PPy was filtered, washed with deionized water followed by acetone and then dried in an oven at 65 °C ± 1 until the total mass became constant.

2.3. Instrumentation

N₂ adsorption–desorption isotherms were measured at 77 K with a BET surface area analyzer “Micrometrics ASAP 2020 gas adsorption apparatus (USA)” using standard continuous procedures. FTIR spectra of adsorbents were recorded using a Perkin Elmer Spectrum 100 FTIR spectrophotometer by a standard KBr technique in the region of 4000–400 cm⁻¹. To analyze the crystal structure of adsorbents, X-ray powder diffraction patterns were recorded using Cu Kα radiation in the 2θ range of 10–80° on a PAN Analytical X'Pert PRO-diffractometer. To evaluate the surface morphology and elemental composition, field emission scanning electron microscopic (FE-SEM) images and energy dispersive X-ray (EDX) spectra of the adsorbents were obtained using a JEOL-JSM 7500F microscope coupled with energy dispersive X-ray. Scanning transmission electron microscopy (STEM) was performed using JEOL-JEM 2100 instrument with an LAB6 filament operated at 200 kV, to measure the particle size and shape. Thermogravimetric analysis was performed using a Perkin-Elmer TGA 4000 by heating the sample from room temperature up to 900 °C under the air atmosphere at a heating rate of 10 °C min⁻¹ to study the thermal behavior of the adsorbents. Malvern Zeta Sizer Nano series (UK) was employed to measure the zeta potentials of the prepared adsorbent.

2.4. Fluoride sorption experiments

Adsorption studies. The adsorption experiments were carried out in batch mode using a thermostatically controlled water bath shaker at 200 rpm for 24 h, by contacting 0.1 g of HTiO₂@PPy with 50 mL of F⁻ solution in plastic bottles. After each adsorption experiment, the adsorbent was filtered from the solution using 0.22 μm filter and the remaining F⁻ ion concentration in the solution was measured in the form of F⁻ conductance using a “Mettler-Toledo” fluoride ion selective electrode, as described elsewhere.^23,24 Typically, a total ionic strength adjustment buffer (Tisab II) was made by dissolving 57 mL glacial acetic acids, 58 g NaCl and 4.0 g of CDTA (cyclo hexylene dinitrile tetraacetate) in 500 mL deionized water; pH was adjusted in between 5.3 and 5.5, using 5 M NaOH and further diluted to 1 L with deionized water.²⁴ TISAB II was then added to the solutions prior to the measurement, to reduce the variation in ionic intensity. The fluoride ion selective electrode was calibrated using 0.1, 1 and 10 mg L⁻¹ standard F⁻ solutions and subsequently fluoride ion concentration of the experimental solution was determined with the help of obtained calibration curve. The effect of pH on the adsorption was examined by keeping 0.1 g of the HTiO₂@PPy in contact with 50 mL F⁻ solutions of 10 mg L⁻¹ concentrations, set at different pH values ranging from 2.5 to 10.5 at temperature 25 (±1.0) °C. Similarly, the effect of adsorbent dosage over F⁻ sorption was carried out by varying the mass of adsorbent (HTiO₂@PPy) from 0.01 g to 0.2 g, at an initial F⁻ concentration of 10 mg L⁻¹, pH 6.5 (±0.2) and temperature 25 (±1.0) °C respectively. The % F⁻ removal efficiency was calculated using eqn (1):where C₀ and C_e are the initial and equilibrium F⁻ concentrations in mg L⁻¹, respectively. Adsorption isotherm experiments were carried out at 15 °C, 25 °C, 35 °C and 45 °C respectively, by contacting 0.1 g of the HTiO₂@PPy with 50 mL F⁻ solutions of different initial concentrations ranged from 10 to 160 mg L⁻¹ at pH 6.5 (±0.2) for 24 h. The adsorption capacity of the adsorbent at equilibrium was calculated using eqn (2):where q_e (mg g⁻¹) is the equilibrium adsorption capacity, m (g) is the mass of adsorbent, and V (L) is the volume of F⁻ solution. To explore the effect of contact time on the F⁻ uptake rate, 1.0 g of the adsorbent was equilibrated with 500 mL F⁻ solutions of initial concentrations 5, 10, 15 and 20 mg L⁻¹, respectively. In brief, aliquots of sample were collected at time zero and then over predetermined time intervals, and the amount of F⁻ ions in the collected samples were analyzed, to evaluate the sorption capacity of adsorbent at time t using eqn (3):where q_t (mg g⁻¹) and C_t (mg L⁻¹) are adsorption capacity and the F⁻ concentration, respectively at time t. To study the effect of temperature on F⁻ uptake rate, kinetic studies were conducted for initial F⁻ concentration of 20 mg L⁻¹ at 15 °C, 25 °C, 35 °C and 45 °C, respectively.

The selectivity of HTiO₂@PPy towards F⁻ ions, was evaluated by conducting the F⁻ adsorption experiments in the presence of known concentrations of anions which commonly occur in natural groundwater. For this NaCl, NaNO₃, Na₂SO₄, NaHCO₃, and Na₂HPO₄·H₂O were added into the F⁻ solution to provide Cl⁻, NO₃⁻, SO₄²⁻, HCO₃⁻ and PO₄³⁻ as competing anions, respectively. Typically, 0.1 g of HTiO₂@PPy was contacted with 50 mL F⁻ solution of 10 mg L⁻¹, containing the ion of interest (Cl⁻, NO₃⁻, SO₄²⁻, HCO₃⁻ and PO₄³⁻ ions) in concentrations 10, 20 and 40 mg L⁻¹ at pH 6.5 (±0.2) and temperature 25 (±1.0) °C. The remaining F⁻ ion concentration was then determined to assess the selectivity of the developed adsorbent.

Adsorbent regeneration. The regeneration and reusability potential of the spent HTiO₂@PPy were tested by contacting 0.1 g of the adsorbent with 50 mL of a F⁻ solution (10 mg L⁻¹) for adsorption at 25 °C and pH 6.5 (±0.2). The adsorbent was then filtered, and the F⁻ loaded adsorbent was then treated with 50 mL of NaOH (0.01–0.1 M) and NH₄OH solutions (0.1–2 M), respectively to desorb the adsorbed F⁻ ions. The % desorption efficiency was then calculated using eqn (4):²⁵where C_des and C_ads are the amounts of total F⁻ desorbed and adsorbed (mg L⁻¹), respectively. After desorption, the exhausted adsorbent was regenerated using 50 mL of 2 M HCl, filtered and re-dispersed in F⁻ solution. Four such consecutive cycles of sorption–desorption were carried out to assess the reusability of HTiO₂@PPy.

3. Results and discussions

3.1. Physico-chemical characterization

BET surface area. The specific surface area which is an essential parameter for the evaluation of textural properties of an adsorbent was calculated using the Brunauer–Emmett–Teller (BET) method. Fig. 1A(a) and (b), respectively demonstrated that N₂ adsorption–desorption isotherms for both HTiO₂ NPs and HTiO₂@PPy belong to type IV based on the IUPAC classification with a clear hysteresis loop which is the characteristic of mesoporous adsorbents (2–50 nm).^26,27 HTiO₂ NPs showed a reasonably high surface area of 295.17 m² g⁻¹ which is in concurrence with what expected for nanostructured materials.²⁸ The BET surface area of HTiO₂@PPy was calculated to be 98.17 m² g⁻¹ which is almost 15 times higher than pure PPy (6.1 m² g⁻¹). This might be attributed to the prevention of π–π* aggregation of the growing PPy chains by incorporation of HTiO₂ NPs in the synthesized HTiO₂@PPy, which caused an increase in surface area of HTiO₂@PPy as compared to pure PPy. The average pore size and pore volume obtained for HTiO₂@PPy were 4.37 nm and 0.11 cm³ g⁻¹, respectively.

Fig. 1 (A) N₂ adsorption–desorption isotherm curves of (a) HTiO₂ NPs and as-synthesized HTiO₂@PPy; (B) zeta potentials of as prepared HTiO₂@PPy at different pH values (a) before and (b) after fluoride adsorption.

Zeta potential. The surface charge of the adsorbent controls the nature of the interaction between adsorbate and adsorbent. Zeta potential measurements as a function of pH were performed to investigate the net surface charge of HTiO₂@PPy before and after fluoride adsorption as demonstrated in Fig. 1B(a) and (b). As shown in Fig. 1B(a), zeta potential values for HTiO₂@PPy gradually decreased from +29.2 mV to −29 mV with increasing pH from 2.5 to 10.5 and point of zero charge i.e. pH_pzc was found to be 8.4. This indicates that, at pH above 8.4, the adsorbent surface is predominately negatively charged whereas, at pH below 8.4, the adsorbent's surface exhibits an overall positive charge and thus can primarily attract anions like F⁻ via electrostatic interactions. Whereas, for fluoride laden HTiO₂@PPy (Fig. 1B(b)), the zeta potential values became less positive and gradually decreased from +8.3 to −17.4. As a result, pH_pzc showed a negative shift towards more acidic pH value i.e. 6.0. This might be attributed to the fact that after F⁻ adsorption, the anticipated electrostatic interaction between F⁻ ions (negatively charged species) and positively charged amine groups of PPy and protonated hydroxyl groups of HTiO₂ NPs, resulted in overall charge neutralization at the surface of adsorbent and eventually as an outcome yielded less positive zeta potential values along with the lowering of pH_pzc of HTiO₂@PPy.

FTIR spectra. The FTIR spectra of as-prepared HTiO₂ NPs as shown in Fig. 2A(a) displays a broad band in the range of 3100–3400 cm⁻¹ due to the stretching vibrations of surface OH groups and molecular H₂O, which confirms the hydrous nature of the synthesized hydrous titanium oxide.^14,22 The band at 1628 cm⁻¹ is due to the bending mode vibration of molecular H₂O, indicating the presence of physisorbed water in HTiO₂ NPs.¹⁶ A prominent peak at 1401 cm⁻¹ is considered as a finger print peak in TiO₂ spectra²⁹ and the bands at 1127 cm⁻¹ and 1054 cm⁻¹ are due to the bending vibrations of the hydroxyl group of metal oxides (Ti–OH) (see the inset graph of HTiO₂ NPs FTIR spectrum from wavenumber 1020 to 1160 cm⁻¹).¹⁶ The band ranged in 760–455 cm⁻¹ is due to the presence of phonon bands of Ti–O–Ti bond in the HTiO₂ NPs.¹⁴ The FTIR spectrum of the as-synthesized HTiO₂@PPy as shown in Fig. 2A(b) demonstrates the peaks at 1458, 1043 and 922 cm⁻¹ which can be assigned to the C–N and C–H stretching vibrations, and the C–H out-of-plane deformation of PPy.³⁰ Moreover, the peaks present at 615 cm⁻¹, 1163 and 1621 cm⁻¹ can be assigned to the Ti–O–Ti bond and bending and stretching vibrations of OH group, respectively.²⁹ The sharp, distinct band present at 3412 cm⁻¹ can be assigned to the OH groups of HTiO₂ NPs as well as N–H stretching peak (3436 cm⁻¹) of PPy. These results confirm the presence of both PPy and HTiO₂ NPs in the as-synthesized HTiO₂@PPy.

Fig. 2 (A) FTIR spectra of (a) HTiO₂ NPs, HTiO₂@PPy (b) before and (c) after F⁻ sorption; (B) XRD patterns of (a) HTiO₂ NPs, HTiO₂@PPy (b) before and (c) after F⁻ sorption.

XRD patterns. The XRD pattern of HTiO₂ NPs (Fig. 2B(a)) shows no distinct peaks for titanium oxide which suggests its amorphous Ti(OH)₄ structure.^14,31,32 The XRD spectra of HTiO₂@PPy (Fig. 2B(b)), showed a broad amorphous diffraction peak in the range of 15° < 2θ < 30° which is the characteristic peak for PPy homopolymer due to the scattering of PPy chains at interplanar spacing.³³ Moreover, no noteworthy alteration noticed in XRD patterns of HTiO₂@PPy after F⁻ sorption (Fig. 2B(c)) suggests that the F⁻ interaction with the nanocomposite has not affected its crystal structure.

SEM and STEM micrographs. SEM micrographs of HTiO₂ NPs and HTiO₂@PPy are shown in Fig. 3A and B, respectively. Fig. 3A shows irregularly shaped agglomerates of nanophase primary particles (average particle size 4–6 nm) of HTiO₂ whereas the SEM image of HTiO₂@PPy (Fig. 3B) shows the formation of globular particles of sizes larger than HTiO₂ NPs. The nucleus effect of HTiO₂ NPs resulted in a homogeneous core–shell type morphology with an encapsulation of the HTiO₂ core by the PPy shell yielding spherical/globular particles of size ranged in between 100 and 150 nm. The STEM image of HTiO₂@PPy after F⁻ adsorption and the corresponding EDS mapping as shown in Fig. 4, provided a direct elemental distribution of C, N, O, Ti and F which confirmed successful incorporation of both PPy and HTiO₂ as well as successful uptake of F⁻ by HTiO₂@PPy. The EDX spectral analysis of HTiO₂@PPy (Fig. 5A(a)) also validated the successful incorporation of both HTiO₂ NPs and PPy in the synthesized nanocomposite by displaying peaks of C, N, Cl (which was incorporated as a dopant from the polymerizing solution), O and Ti at 0.277, 0.392, 2.622, 0.525 and 4.511 keV, respectively. After F⁻ sorption, the spectrum for HTiO₂@PPy (Fig. 5A(b)) displayed an additional peak of F which approves the F⁻ adsorption by HTiO₂@PPy. Moreover, it was also noticed that fluoride laden HTiO₂@PPy (Fig. 5A(b)) exhibited a smaller peak of Cl at 2.622 keV than pristine HTiO₂@PPy (Fig. 5A(a)), which suggests that F⁻ ions might have replaced doped Cl⁻ ions present in HTiO₂@PPy via the ion-exchange mechanism. To support the ion-exchange mechanism, we have performed the silver nitrate (AgNO₃) solution test with treated water. Typically, a 10 mg L⁻¹ fluoride solution was made with a subsequent dilution of fluoride stock solution with de-ionized water. The pH of above fluoride solution was noted to be 6.7 (no pH adjustments were done). After that, a characteristic batch study was conducted using 0.1 g of adsorbent (HTiO₂@PPy) and 50 mL of 10 mg L⁻¹ fluoride solution. A control experiment was also conducted using 50 mL of 10 mg L⁻¹ fluoride solution with no adsorbent i.e. HTiO₂@PPy. In another control experiment, we have added 0.1 g of HTiO₂@PPy into 50 mL deionized water. Then both controls and HTiO₂@PPy treated fluoride solution were filtered, and the filtrates were treated with aq. AgNO₃ solution. It was observed that white colored precipitate of silver chloride (AgCl) appeared in the HTiO₂@PPy treated fluoride solution whereas no such precipitate formed in the control experiments. This observation justified the presence of Cl⁻ ions in the filtrate obtained from HTiO₂@PPy treated fluoride solution that might have been replaced as a result of ion-exchange of doped Cl⁻ ions present in the polymer with F⁻ ions present in the fluoride solution.

Fig. 3 Scanning electron microscopic (SEM) images of (A) HTiO₂ NPs (100k×) and (B) HTiO₂@PPy (25k×).

Fig. 4 Scanning transmission electron microscopic (STEM) images of HTiO₂@PPy after F⁻ sorption with EDS mapping of carbon (C), nitrogen (N), oxygen (O), titanium (Ti) and fluorine (F).

Fig. 5 (A) EDX spectra of HTiO₂@PPy (a) before and (b) after F⁻ adsorption and (B) TGA plots of (a) HTiO₂ NPs and (b) HTiO₂@PPy.

Thermogravimetric analysis. TGA plots obtained for HTiO₂ NPs, and HTiO₂@PPy are shown in Fig. 5B(a) and (b), respectively. Fig. 5B(a) showed an initial weight loss of 8.2% up to 100 °C which corresponds to the loss of physically attached water molecules,²² thereafter the weight loss of 11.5% up to 400 °C is mainly due to the loss of water from the hydroxyl groups²² and later above 400 °C, the curve eventually flattened. On the contrary, Fig. 5B(b) showed an initial weight loss of 12.2% up to 200 °C which is attributed to the liberation of entrapped water in the HTiO₂@PPy indicating the hydrous³⁴ nature of HTiO₂ NPs. The additional major weight loss of 59.8% between 200 °C and 580 °C mainly corresponds to the thermal decomposition of organic PPy chains.³⁵ Therefore, the developed HTiO₂@PPy can be effectively used for water treatment applications in open air atmospheres for temperatures up to 200 °C.

3.2. Adsorption characteristics

% F⁻ removal by HTiO₂@PPy, pristine HTiO₂ NPs and pure PPy. The F⁻ removal efficiency of pristine HTiO₂ NPs, pure PPy and as-synthesized HTiO₂@PPy, using 0.05 g of each adsorbent and 50 mL of a 30 mg L⁻¹ F⁻ solution at pH 6.5 (±0.2) and temperature 25 (±1.0) °C was compared and shown in Fig. S1 (ESI†). It was observed that as-synthesized HTiO₂@PPy showed significantly greater F⁻ removal efficiency than both PPy and HTiO₂ NPs. It might be possible that F⁻ adsorbing properties of PPy and HTiO₂ NPs, along with prevention of aggregation of PPy chains due to the incorporation of HTiO₂ NPs, all have synergistically resulted in improved F⁻ adsorptive performance of HTiO₂@PPy as compared to the base material counterparts viz. PPy and HTiO₂ NPs.

Effect of pH. pH is a critical parameter that evaluates not only the adsorptive potential of an adsorbent but also gives an insight into underlying mechanisms governing the adsorption processes.³⁶ Fig. 6A demonstrates the impact of solution pH over the F⁻ adsorption efficiency of HTiO₂@PPy, whereby adsorption increased with increase in pH reaching a maximum at pH 3.5, thereafter remained nearly constant up to pH 8.5 and then decreased abruptly beyond pH 8.5. This observation can be well elucidated by the change in the adsorbent's surface charge as demonstrated in zeta potential plot (Fig. 1B(a)). A diminished F⁻ removal efficiency observed in strongly acidic medium, might be attributed to the formation of hydrofluoric acid (HF), due to the shift in the equilibrium of eqn (5) more towards left as a result of the common ion effect.³⁷

HF_(aq) ⇌ H_(aq)⁺ + F_(aq)⁻ (5)

Fig. 6 Effect of (A) initial solution pH and (B) adsorbent dosages on % F⁻ removal by HTiO₂@PPy.

At pH values lower than pH_pzc ≈ 8.4, the protonation of nitrogen groups of PPy and surface hydroxyl groups of the HTiO₂ NPs gave rise to an overall positive charge on the adsorbent's surface which resulted in the adsorption of negatively charged F⁻ ions by electrostatic interactions.^17,38 The F⁻ sorption at pH 8.5 ≈ pH_pzc signifies that the sorption has also occurred on the electrically neutral surface of the adsorbent which might be due to the possible ion exchange between small sized F⁻ ions and Cl⁻ ions incorporated in the PPy chain as dopants.¹⁷ This indicates the prevalence of an ion exchange mechanism as well, along with the electrostatic interaction in the overall F⁻ sorption onto the HTiO₂@PPy.

Furthermore, the predominantly negatively charged surface of HTiO₂@PPy (Fig. 1B) at pH above 8.5 > pH_pzc, due to the deprotonation of NH groups and surface OH groups present in the nanocomposite,³⁹ resulted in a sharp decline in the F⁻ removal efficiency of HTiO₂@PPy as a consequence of the electrostatic repulsion between anionic F⁻ ions and the negatively charged adsorbent's surface.⁴⁰ Moreover, OH⁻ ions present in highly alkaline conditions being isoelectronic in nature, may also compete with the F⁻ ions for the same active sites on the adsorbent's surface.⁴¹ Hence, HTiO₂@PPy exhibited a commendable defluoridation efficiency over a wide pH range of 3.5–8.5 and hence, pH 6.5 (±0.2) which characteristically lies in the drinking water pH range was chosen as an optimum pH for further experiments.

Effect of adsorbent dosage. The F⁻ removal efficiency of HTiO₂@PPy as a function of the adsorbent dose was studied to optimize the minimum mass required for the maximum F⁻ removal, and the result is shown in Fig. 6B. The F⁻ removal efficiency increased from 28.90% to 95.98% as the HTiO₂@PPy dose was increased from 0.01 to 0.1 g due to the increase in the number of active sites available for F⁻ sorption.⁴² Thereafter, the F⁻ removal efficiency remained nearly constant with a very less proportionate increase of ∼1.4% on increasing the adsorbent dose from 0.1 g to 0.2 g. The saturation plateau in F⁻ removal efficiency of HTiO₂@PPy (Fig. 6B) can be attributed to the more frequent collisions between adsorbent particles at higher dosages which eventually lead to particle aggregation and result in a decrease in the overall effective surface area of adsorbent.⁴³ Therefore, 0.1 g of adsorbent dosage was chosen as the optimal adsorbent dosage for the subsequent adsorption studies.

Adsorption kinetics. The adsorption kinetics is vital for the selection of best operating conditions if real field water treatment application needs to be explored and also gives an important idea about the adsorbate uptake rate and equilibrium time.⁴⁴ As shown in Fig. 7A, F⁻ adsorption by HTiO₂@PPy exhibited a rapid initial uptake up to 5 min; after that removal rate slowed, reaching equilibrium in 5, 10, 15 and 30 min, for the initial F⁻ concentrations of 5, 10, 15 and 20 mg L⁻¹, respectively. The rapid uptake in the first 5 min, can be explained by the small mass transfer resistance which leads to the faster occupation of vacant active sites available at the adsorbent's surface by F⁻ ions. Thereafter, the covering of the nanocomposite's surface by F⁻ ions, and the slow diffusion of F⁻ ions into the internal surfaces slowed down the overall adsorption rate, finally resulting in equilibrium. The experimental data was further analyzed using pseudo-first-order and pseudo-second-order kinetic models represented by eqn (6) and (7), respectively:where q_e and q_t are adsorption capacities in mg g⁻¹ at equilibrium and time t (min), respectively; k₁ (min⁻¹) and k₂ (g mg⁻¹ min⁻¹) are the pseudo-first-order and pseudo-second-order rate constants, respectively.

Fig. 7 Non-linear pseudo-first-order and pseudo-second-order kinetic modelling for effect of contact time over F⁻ adsorption at (A) different initial concentrations and (B) different temperatures.

The non-linearized forms of the pseudo-first-order and pseudo-second-order kinetic models can be represented by eqn (8) and (9), respectively:

q_t = q_e(1 − exp(−k₁t)) (8)

Fig. 7A shows the fits of the kinetic data to the non-linear pseudo-second-order and pseudo-first-order, kinetic models. The kinetic parameters calculated from the non-linear regression analysis of the plots (Fig. 7A) are presented in Table 1. It was noticed that although the pseudo-first-order kinetic model gave satisfactorily good correlation coefficient values (R² > 0.900); yet the best fit was obtained using pseudo-second-order kinetic model with higher correlation coefficient values (R² > 0.990). The q_e values obtained from the pseudo-second-order model were also close and consistent with the experimental q_e values. The pseudo-second-order rate constant k₂ decreased with an increase in the initial F⁻ concentration as shown in Table 1 which indicates towards the decrease in F⁻ uptake rate with an increase in the initial F⁻ concentration.

Table 1 Kinetic parameters for F⁻ sorption by HTiO₂@PPy at different initial concentrations^a

Kinetic models	Initial concentrations (mg L^−1)
	5	10	15	20
a Units: qe: (mg g^−1), k1: (min^−1), k2: (g mg^−1 min^−1), kint: (mg g^−1 min^−1)^1/2, Ci: (mg g^−1).
Pseudo-first-order non-linear
k1	3.131	2.592	2.113	1.938
qe	2.419	4.738	6.864	8.973
R^2	0.992	0.994	0.993	0.988
sy.x	0.018	0.076	0.139	0.279
Pseudo-second-order non-linear
k2	8.250	2.257	0.996	0.597
qe	2.429	4.772	6.931	9.087
R^2	0.999	0.998	0.998	0.994
sy.x	0.005	0.032	0.052	0.161
Intraparticle diffusion model
kint	0.054	0.123	0.399	0.491
Ci	2.275	4.316	5.773	7.429
R^2	0.826	0.961	0.915	0.883

---

Apart from the adsorption at the exterior surface of adsorbent, adsorbate molecules may also diffuse into the interior of adsorbent. Thus, to identify the rate-determining step⁴⁵ for the F⁻ adsorption by HTiO₂@PPy, the kinetic data was further tested using an intraparticle diffusion model proposed by Weber and Morris which is expressed as eqn (10):

q_t = k_intt^0.5 + C_i (10)

where k_int (mg g⁻¹ min^−0.5) is the intra-particle diffusion rate constant, and C_i (mg g⁻¹) is the intercept related to the boundary layer thickness. Fig. S2 (ESI†) shows intraparticle diffusion plots for the initial F⁻ concentrations 5, 10, 15 and 20 mg L⁻¹ respectively. An adsorption process usually proceeds in three phases: (1) in the first phase, adsorbate diffuses from the bulk solution to external surface of adsorbent, i.e. boundary layer diffusion; (2) during the second phase, a slower intraparticle diffusion and (3) subsequently, the final phase of equilibrium stage of adsorption process.⁴⁶ Multilinearity of the intraparticle diffusion plots indicates that F⁻ sorption by HTiO₂@PPy is a complex process in which both boundary layer and intraparticle diffusion determines the rate of the overall adsorption process. As shown in Fig. S2 (ESI†), due to very fast kinetics, at lower F⁻ concentrations (5 and 10 mg L⁻¹) final equilibrium stage is merged with intraparticle pore diffusion^21,34 whereas higher concentrations (15 and 20 mg L⁻¹) show distinct three linear sections. The values of k_int and C_i were calculated from slopes and intercepts of the first linear region and are presented in Table 1. The R² values ≈0.9, signify that the intraparticle diffusion process plays a significant role in F⁻ sorption by HTiO₂@PPy.

The effect of temperature and time simultaneously on the rate of F⁻ sorption by HTiO₂@PPy was also studied. It was noted that q_e increased with increase in temperature thereby suggesting an endothermic nature of adsorption. The experimental data obtained was further analyzed using pseudo-first-order and pseudo-second order non-linear eqn (8) and (9), and respective plots are shown in Fig. 7B. The kinetic parameters obtained from non-linear regression analysis of the Fig. 7B are presented in Table 2. The values of regression coefficients (R²) at any given temperature indicated that the pseudo-second-order model described the temperature dependent kinetic data better than pseudo-first-order model. The pseudo-second-order rate constant k₂ increased from 0.469 to 0.958 g mg⁻¹ min⁻¹ with an increase in temperature from 15 °C to 45 °C that can be attributed to the Arrhenius law of temperature dependence of reaction rates.⁴⁷ An incredibly fast kinetics for F⁻ sorption by HTiO₂@PPy as compared to most other literature reported adsorbents^{17,19–21,48–52} was comparatively assessed and presented in Table S1 (ESI†).

Table 2 Kinetic parameters evaluated for F⁻ sorption by HTiO₂@PPy at different temperatures

Temp. (°C)	Pseudo-first-order model			Pseudo-second-order model
	k1 (min^−1)	qe (mg g^−1)	R^2	k2 (g mg^−1 min^−1)	qe (mg g^−1)	R^2
15	1.657	8.080	0.981	0.469	8.265	0.994
25	2.040	8.867	0.989	0.657	9.013	0.995
35	2.410	9.100	0.994	0.931	9.211	0.999
45	2.432	9.340	0.995	0.958	9.446	0.998

---

Adsorption isotherm. To understand the F⁻ adsorptive performance of HTiO₂@PPy, the adsorption isotherms for F⁻ sorption were obtained at 15 °C, 25 °C, 35 °C and 45 °C, respectively and displayed in Fig. 8A. It was observed that the adsorption capacity increased with an increase in the initial F⁻ concentration and reached a constant value in the high initial concentration range. To quantitatively analyze the adsorption isotherm data, two most widely used isotherm models, viz. the Langmuir and Freundlich were employed. The Langmuir isotherm model can be expressed in non-linear and linear forms by eqn (11) and (12) respectively:where, q_max (mg g⁻¹) is the Langmuir's maximum adsorption capacity and b (L mg⁻¹) is the Langmuir constant. The favorability of an adsorption process is described by the dimensionless separation factor R_L and can be expressed as eqn (13):where the significance of C₀ and b are already specified earlier. The value of R_L presents the tendency of the isotherm to be favorable (0 < R_L < 1), linear (R_L = 1), unfavorable (R_L > 1), or irreversible (R_L = 0).⁵³ Besides, smaller R_L value indicates greater affinity between the adsorbate and adsorbent. Fig. 8A and B display the Langmuir non-linear and linear plot at different temperatures, respectively.

Fig. 8 F⁻ sorption isotherm data fit with (A) non-linear Langmuir and Freundlich isotherm models and (B) linear Langmuir model.

The Freundlich isotherm can be expressed in the non-linear and linear forms by eqn (14) and (15) respectively:

q_e = K_fC_e^1/n (14)

where K_f (mg g⁻¹) represents the Freundlich constant, and 1/n is the surface heterogeneity factor of the adsorbent. Fig. 8A and S3 (ESI†) show the Freundlich non-linear and linear plot respectively, for adsorption isotherm data. The relative Langmuir and Freundlich parameters calculated and presented in Table 3 showed that, though the correlation coefficient (R²) values obtained for Freundlich model fit were satisfactorily good (R² > 0.900), but still the Langmuir model fit gave a better correlation coefficient values (R² > 0.970) for both non-linear and linear regression analysis of the isotherm data. Hence, the F⁻ sorption by HTiO₂@PPy can be better described by the Langmuir model which assumes monolayer adsorption at the homogeneous surface than the Freundlich model which describes non-ideal adsorption on the heterogeneous surface.⁵⁴ The Langmuir's maximum adsorption capacity q_max evaluated from linear and non-linear analysis of isotherm data came out to be 31.64 and 31.93 mg g⁻¹ respectively, at 25 °C. The R_L values at different initial F⁻ concentrations (Table 3) for HTiO₂@PPy were in between 0 and 1, indicating that the developed adsorbent had a high affinity towards F⁻.⁹

Table 3 Langmuir and Freundlich isotherm parameters for F⁻ sorption by HTiO₂@PPy^a

Isotherm models	Temperatures (K)
	15 °C	25 °C	35 °C	45 °C
a Units: qmax: (mg g^−1), b: (L mg^−1), Kf: (mg g^−1).
Langmuir linear
qmax	26.316	31.645	33.333	34.364
b	0.123	0.217	0.210	0.262
RL	0.448	0.316	0.322	0.276
R^2	0.991	0.998	0.998	0.999
Non-linear
qmax	26.06	31.93	33.32	34.42
b	0.113	0.183	0.184	0.227
R^2	0.935	0.973	0.970	0.973
sy.x	2.001	1.706	1.905	1.901
Freundlich linear
Kf	6.247	7.663	8.030	8.819
1/n	0.310	0.334	0.339	0.332
R^2	0.994	0.958	0.960	0.953
Non-linear
Kf	6.817	9.941	10.23	11.40
1/n	0.287	0.263	0.268	0.255
R^2	0.985	0.937	0.953	0.940
sy.x	0.969	2.598	2.372	2.790

---

The F⁻ removal efficiency of HTiO₂@PPy can not be directly weighed against other reported adsorbents because of the different applied experimental conditions but however it was observed that a significantly reasonable F⁻ adsorption capacity was achieved near neutral pH by HTiO₂@PPy in comparison to the most of the other adsorbents (Table 4).^{16–21,29,32,55,56} Hence, it is anticipated that HTiO₂@PPy can efficiently defluoridate drinking water without prior pH adjustments.

Table 4 Comparison of F⁻ adsorption capacity of HTiO₂@PPy with other adsorbents

Adsorbents	qmax (mg g^−1)	Optimum pH	References
Fe doped titanium oxide	53.22	7	32
TiO2, Ti–Ce and Ti–La	1.8, 9.6 and 15.1	6	16
Fe–Ti oxide nanoadsorbent	47	na	29
PPy	6.37	na	17
PPy/chitosan composite	6.7	na	18
PPy/alumina composite	8	na	19
PPy/Fe3O4 nanocomposite	17.6–22.3	6.5	20
PPy/HSnO nanocomposite	26.16	6.5 ± 0.1	21
Hydrous bismuth oxides (HBO)	1.93	7.0 ± 0.1	55
Tea–Al–Fe	18.52	7.0 ± 0.2	56
HTiO2@PPy	31.93	6.5 ± 0.2	Present work

---

Thermodynamic and activation parameters. Thermodynamic parameters, associated with an adsorption process i.e. the standard Gibbs free energy change (ΔG°), standard enthalpy change (ΔH°) and standard entropy change (ΔS°) were calculated from adsorption isotherm data by applying the Van't Hoff's eqn (16) and (17):where, R is the gas constant (0.008314 kJ mol⁻¹ K⁻¹), T is the temperature (K), K_c is the thermodynamic equilibrium constant, m is the adsorbent dose (g L⁻¹), and the ratio mq_e/C_e is the adsorption affinity.

Accordingly, the values of ΔH° and ΔS° were evaluated from the slope and intercept of Van't Hoff plot (Fig. 9A) and the thermodynamic constants obtained are presented in Table 5. The positive value of ΔH° confirms the endothermic nature of adsorption process.⁵⁷ Furthermore, ΔH° value 27.80 kJ mol⁻¹ < 80 kJ mol⁻¹ signifies the physical nature of F⁻ adsorption by HTiO₂@PPy.⁵⁸ The positive value of ΔS° implies to the increased disorder in the system with the changes in the hydration state of adsorbed F⁻ ions.⁵⁹ The decreasingly negative values of ΔG° with the increase in temperature indicates the spontaneous nature of F⁻ adsorption.⁵⁷

Fig. 9 (A) Plot of ln(mq_e/C_e) versus 1/T: Vant Hoff's plot and (B) plot of ln k₂ versus 1/T: Arrhenius plot.

Table 5 Thermodynamic and activation parameters for F⁻ adsorption by HTiO₂@PPy

Temperature (K)	ΔH° (kJ mol^−1)	ΔS° (kJ mol^−1 K^−1)	ΔG° (kJ mol^−1)	Ea (kJ mol^−1)	A (g mg^−1 min^−1)
15	27.80	0.113	−4.52	31.80	1.05 × 10^5
25			−6.10
35			−6.68
45			−7.22

---

Furthermore, to determine the activation parameters of the present adsorption system, the Arrhenius eqn (18) was adopted using the pseudo-second-order rate constants, k₂ at different temperatures.

where A is the temperature-independent pre-exponential factor (g mg⁻¹ min⁻¹); E_a is the activation energy (kJ mol⁻¹) of the adsorption, R and T have their usual significances as mentioned earlier. The correlation coefficient obtained (R² = 0.949) from the plot of ln k₂ versus 1/T as shown in Fig. 9B indicates that the F⁻ sorption onto HTiO₂@PPy follows Arrhenius law well. The adsorption activation energy i.e. E_a obtained for the present adsorption system (Table 5) was evaluated as 31.80 kJ mol⁻¹ which is lower than 40 kJ mol⁻¹ and hence confirming the physical nature of F⁻ sorption by HTiO₂@PPy.^60,61

Effect of co-existing anions. Anions such as chloride (Cl⁻), nitrate (NO₃⁻), bicarbonate (HCO₃⁻), sulfate (SO₄²⁻) and phosphate (PO₄³⁻) commonly exist in the surface water or groundwater and could compete with F⁻ ions for the active adsorption sites during the adsorption process. Hence, the influence of all these anions on F⁻ adsorption by HTiO₂@PPy NC was experimentally investigated as shown in Fig. 10A. Cl⁻ and NO₃⁻ being low-affinity ligands involve weaker bonds via outer-sphere complex formation with the active sites and thus rarely affects the F⁻ adsorption. Fig. 10A showed that F⁻ sorption slightly increased in the presence of Cl⁻ and NO₃⁻, as reported in literature some anions would enhance columbic repulsive forces, and some would compete with F⁻ for the active sites.⁴⁵ SO₄²⁻ showed negligible effect on F⁻ sorption for the studied concentration range. The presence of HCO₃⁻ slightly affected the F⁻ sorption at 40 mg L⁻¹ concentration causing a ∼3% decrease in fluoride removal efficiency. The slight influence of bicarbonate ions at the studied concentration range, (even when bicarbonate ions get hydrolyzed⁵ and might increase the solution pH up to 8.0) can be further justified by the fact that, the developed adsorbent HTiO₂@PPy has shown nearly constant fluoride removal efficiency in the range of pH 3.5 to 8.5 as discussed in Section 3.2 and shown in Fig. 6A. Hence, it can be expected to observe no significant change in fluoride removal efficiency of HTiO₂@PPy when pH is raised from 6.5 to 8.0–8.5. PO₄³⁻ ions showed the highest negative impact on the F⁻ removal and reduced the F⁻ removal efficiency of HTiO₂@PPy by 8.22% at the concentration of 40 mg L⁻¹ which was consistent with previous studies.^62,63 This may be due to the adsorption of PO₄³⁻ ions onto specific adsorption sites on the HTiO₂@PPy surface, different from the F⁻ adsorption sites. However, the concentration of PO₄³⁻ ion in natural water is much lower (0–5 mg L⁻¹)⁶⁴ and hence, practically not much interference in natural environments would be caused by PO₄³⁻ ions. Hence, it can be concluded that HTiO₂@PPy shows a good selectivity towards the F⁻ removal in the presence of other groundwater coexisting anions.

Fig. 10 (A) Effect of coexisting anions on the F⁻ removal by HTiO₂@PPy and (B) F⁻ removal efficiency of the recycled HTiO₂@PPy using 1.0 M NH₄OH and 0.05 M NaOH.

Real field study. The potential applicability of the as-prepared HTiO₂@PPy for the removal of F⁻ from the real field (groundwater) was demonstrated by collecting a groundwater sample from a borehole of NW province of South Africa and treating it with HTiO₂@PPy. Some physicochemical parameters of collected groundwater sample were estimated as pH (6.67), electrical conductivity (670 μS cm⁻¹), F⁻ (1.12 mg L⁻¹), Cl⁻ (19.27 mg L⁻¹), NO₃⁻ (1.32 mg L⁻¹), SO₄²⁻ (14.28 mg L⁻¹) and PO₄³⁻ (1.05 mg L⁻¹). The collected groundwater sample was spiked with F⁻ to increase its concentration to 5.1 mg L⁻¹ and then treated with HTiO₂@PPy (0.05 to 0.3 g/50 mL) at 25 (±1.0) °C, pH 7.4 for 24 h in the batch mode. The relationship between the residual F⁻ concentration and adsorbent dosage along with the % F⁻ removal is shown in Fig. S4 (ESI†). It was noted that 0.1 g of HTiO₂@PPy, effectively reduced the F⁻ concentration to 0.97 mg L⁻¹ which is below the standard permissible limit of F⁻ (<1.5 mg L⁻¹)⁹ in the drinking water.

3.3. Desorption studies

The effective regeneration and reusability are desirable properties of an adsorbent so that it can be put into cyclic use in a cost effective manner for its real field application. Fig. S5(A) and (B) (ESI†) show the results obtained for the desorption of adsorbed F⁻ ions from 0.1 g HTiO₂@PPy carried out using NaOH (0.01–0.1 M) and NH₄OH (0.2–2 M) solutions, respectively. It was noted that the highest desorption efficiency achieved using 0.1 M NaOH solution was 87.21% (Fig. S5(A)†) whereas 96.56% of desorption efficiency was achieved using 1 M NH₄OH solution (Fig. S5(B)†). The percentage of F⁻ desorbed appeared to be very low in the case of NaOH solutions which might be due to the destruction of polymer backbone in the as-synthesized HTiO₂@PPy by NaOH (from the preliminary studies, it was found that even the lower concentrations of NaOH solutions were producing pale yellow color after desorption) whereas NH₄OH showed excellent desorption efficiency even at the lower concentrations. A comparative regeneration and reusability studied carried out for F⁻ removal by HTiO₂@PPy using 0.05 M NaOH and 1.0 M NH₄OH as shown in Fig. 10B indicated that HTiO₂@PPy could be more efficiently recycled by 1 M NH₄OH in comparison to 0.05 M NaOH.

3.4. Mechanism of fluoride sorption

The basic mechanism of F⁻ adsorption using various adsorbents is primarily based on the electrostatic attraction and ion exchange reaction, which in turn depends on the pH of the medium. Scheme 1 hypothetically explains the prime two possible mechanisms of F⁻ adsorption by HTiO₂@PPy as: (a) at acidic pH, protonated hydroxyl groups (OH₂⁺) groups of HTiO₂ NPs and protonated amine groups (NH⁺) groups of PPy might electrostatically interact with negatively charged F⁻ ions.²¹ To explore the electrostatic interaction of F⁻ ions with positively charged functional groups of adsorbent, FTIR spectra of HTiO₂@PPy before (Fig. 2A(b)) and after F⁻ adsorption (Fig. 2A(c)) were critically analyzed. Major spectral variations after F⁻ sorption were observed at 3416, 1172 and 1621 cm⁻¹. The spectrum of HTiO₂@PPy after F⁻ adsorption (Fig. 2A(c)) showed a slight shift and increase in the intensity of the peak at 3412 cm⁻¹ towards a higher wavenumber of 3416 cm⁻¹ which can be attributed to the electrostatic interaction of F⁻ ions with NH groups of PPy and OH groups of HTiO₂ NPs. Moreover, the peak contributed by bending vibrations of OH groups of HTiO₂@PPy (Fig. 2A(b)) at 1163 cm⁻¹, also exhibited a slight shift towards the higher wavenumber of 1172 cm⁻¹ in FTIR spectrum of HTiO₂@PPy after F⁻ sorption (Fig. 2A(c)), indicating an ionic interaction of OH groups with F⁻ ions. An observed increase in the intensity of the peak at 1621 cm⁻¹ in Fig. 2A(c), further indicates the interaction of F⁻ ions with surface OH groups of HTiO₂@PPy. All above spectral findings suggest the involvement of both protonated NH groups and OH groups of HTiO₂@PPy in the F⁻ adsorption via electrostatic interaction. (b) At neutral pH, there might be an ion exchange between the Cl⁻ ions doped in PPy chains and F⁻ ions present in the solution. The involvement of ion-exchange has already been discussed in Section 2.3 in the EDX analysis of HTiO₂@PPy before and after fluoride sorption. Thus, conclusively the overall F⁻ adsorption mechanism onto HTiO₂@PPy involves both electrostatic interactions and ion-exchange.

Scheme 1 A possible mechanism for F⁻ adsorption on as-synthesized HTiO₂@PPy.

4. Conclusion

Hydrous TiO₂@polypyrrole hybrid nanocomposite was synthesized via in situ chemical oxidative polymerization technique, characterized and applied for water defluoridation. A substantially higher surface area of 98.17 m² g⁻¹, larger pore size and pore volume of 4.37 nm and 0.11 cm³ g⁻¹, respectively were obtained. The wider pH range (3.5 to 8.5) applicability was achieved due to a high pH_pzc value i.e. 8.4, making it effective even in a natural water environment. HTiO₂@PPy displayed very fast kinetics which followed pseudo-second-order kinetic model (R² = 0.994–0.999; sy.x = 0.005–0.161) with attainment of equilibrium in only 5–30 min and significantly high pseudo-second-order rate constants (k₂) of 8.250, 2.257, 0.996 and 0.597 g mg⁻¹ min⁻¹, for initial F⁻ concentrations 5, 10, 15 and 20 mg L⁻¹, respectively. The adsorption activation energy, calculated using kinetic data was evaluated as 31.80 kJ mol⁻¹ and suggested physical nature of F⁻ adsorption onto HTiO₂@PPy. The sorption isotherm data was well described by the Langmuir model (R² = 0.973; sy.x = 1.706) with maximum adsorption capacity of 31.93 mg g⁻¹ at 25 °C. Thermodynamic parameters, evaluated using isotherm data, were determined as ΔH° = 27.80 kJ mol⁻¹, ΔS° = 0.113 kJ mol⁻¹ K⁻¹ and ΔG° = −4.52 to −7.22 kJ mol⁻¹ and confirmed the endothermic and spontaneous nature of adsorption process. The effect of coexisting anions showed that the fluoride removal efficiency of HTiO₂@PPy was reduced by ∼3% and 8.22% in the presence of 40 mg L⁻¹ concentration of HCO₃⁻ ions and PO₄³⁻ ions, respectively. Considering the typical concentration ranges of coexisting anions in natural groundwater, the effect of PO₄³⁻ ions is not practically beneficial and hence can be overlooked. Regeneration tests demonstrated that 1 M NH₄OH can optimally desorb F⁻ ions from F⁻ laden HTiO₂@PPy and the regenerated adsorbent can be used for three adsorption–desorption cycles without much loss in its F⁻ removal efficiency. The real field studies demonstrated that HTiO₂@PPy can be effectively used for defluoridation of fluoride containing groundwater. The possible underlying mechanisms of F⁻ sorption onto HTiO₂@PPy were interpreted to be electrostatic interactions and ion exchange. Hence, if compared with our previous study in which we studied the defluoridation potential of a PPy based nanocomposite using hydrous tin oxide (PPy/HSnO NC 3),²¹ it can be easily concluded that potentially better fluoride adsorption results were obtained from HTiO₂@PPy and it is well expected to make a potential adsorbent for defluoridation of drinking water.

Acknowledgements

The research was supported by National Research Foundation (NRF) (Grant No. 90742), Water Research Commission (WRC) (Grant No. 1003392) and Eskom, South Africa. The authors thank the Council for Scientific and Industrial Research (CSIR), South Africa and characterization unit at the DST-CSIR National Centre for Nanostructured Materials for all their contribution to this project along with the assistance with materials characterization. The authors like to extend the gratitude to Prof. L. Chimuka (Wits University, SA) for assistance with Ion Chromatography.

References

1. V. Chaudhary and S. Prasad, Anal. Methods, 2015, 7, 8304–8314.
2. W. Tang, P. Kovalsky, B. Cao and T. D. Waite, Water Res., 2016, 88, 112–121.
3. Y. Yu, C. Wang, X. Guo and J. Paul Chen, J. Colloid Interface Sci., 2015, 441, 113–120.
4. S. V. Jadhav, E. Bringas, G. D. Yadav, V. K. Rathod, I. Ortiz and K. V. Marathe, J. Environ. Manage., 2015, 162, 306–325.
5. K. Zhang, S. Wu, J. He, L. Chen, X. Cai, K. Chen, Y. Li, B. Sun, D. Lin, G. Liu, L. Kong and J. Liu, J. Colloid Interface Sci., 2016, 475, 17–25.
6. A. Bhatnagar, E. Kumar and M. Sillanpää, Chem. Eng. J., 2011, 171, 811–840.
7. P. Loganathan, S. Vigneswaran, J. Kandasamy and R. Naidu, J. Hazard. Mater., 2013, 248–249, 1–19.
8. S. Jagtap, M. K. Yenkie, N. Labhsetwar and S. Rayalu, Chem. Rev., 2012, 112, 2454–2466.
9. L. H. Velazquez-jimenez, R. H. Hurt, J. Matos and J. R. Rangel-mendez, Environ. Sci. Technol., 2014, 48, 1166–1174.
10. K. Zhang, S. Wu, X. Wang, J. He, B. Sun, Y. Jia, T. Luo, F. Meng, Z. Jin, D. Lin, W. Shen, L. Kong and J. Liu, J. Colloid Interface Sci., 2015, 446, 194–202.
11. B. Pan, B. Pan, W. Zhang, L. Lv, Q. Zhang and S. Zheng, Chem. Eng. J., 2009, 151, 19–29.
12. E. H. Borai, M. M. E. Breky, M. S. Sayed and M. M. Abo-Aly, J. Colloid Interface Sci., 2015, 450, 17–25.
13. S. P. Suriyaraj and R. Selvakumar, RSC Adv., 2016, 6, 10565–10583.
14. J. Liu, W. Li, Y. Liu, Q. Zeng and S. Hong, Appl. Surf. Sci., 2014, 293, 46–54.
15. T. Wajima, Y. Umeta, S. Narita and K. Sugawara, Desalination, 2009, 249, 323–330.
16. Z. Li, S. Deng, X. Zhang, W. Zhou, J. Huang and G. Yu, Front. Environ. Sci. Eng. China, 2010, 4, 414–420.
17. M. Karthikeyan, K. K. Satheeshkumar and K. P. Elango, J. Hazard. Mater., 2009, 167, 300–305.
18. M. Bhaumik, T. Y. Leswifi, A. Maity, V. V. Srinivasu and M. S. Onyango, J. Hazard. Mater., 2011, 186, 150–159.
19. M. Karthikeyan, K. K. S. Kumar and K. P. Elango, Desalination, 2011, 267, 49–56.
20. M. Karthikeyan, K. K. Satheesh Kumar and K. P. Elango, J. Fluorine Chem., 2009, 130, 894–901.
21. K. Parashar, N. Ballav, S. Debnath, K. Pillay and A. Maity, J. Colloid Interface Sci., 2016, 476, 103–118.
22. S. Debnath and U. C. Ghosh, Chem. Eng. J., 2009, 152, 480–491.
23. J. Zhang, S. Xie and Y. Ho, J. Hazard. Mater., 2009, 165, 218–222.
24. V. Ganvir and K. Das, J. Hazard. Mater., 2011, 185, 1287–1294.
25. M. G. Sujana, A. Mishra and B. C. Acharya, Appl. Surf. Sci., 2013, 270, 767–776.
26. S. Wu, K. Zhang, J. He, X. Cai, K. Chen, Y. Li, B. Sun, L. Kong and J. Liu, J. Colloid Interface Sci., 2016, 464, 238–245.
27. K. Kaneko, J. Membr. Sci., 1994, 96, 59–89.
28. S. Debnath, D. Nandi and U. C. Ghosh, J. Chem. Eng. Data, 2011, 56, 3021–3028.
29. L. Chen, B. He, S. He, T. Wang, C. Su and Y. Jin, Powder Technol., 2012, 227, 3–8.
30. M. Bhaumik, S. Agarwal, V. K. Gupta and A. Maity, J. Colloid Interface Sci., 2016, 470, 257–267.
31. H. Tel, Y. Altas and M. S. Taner, J. Hazard. Mater., 2004, 112, 225–231.
32. L. Chen, S. He, B. Y. He, T. J. Wang, C. L. Su, C. Zhang and Y. Jin, Ind. Eng. Chem. Res., 2012, 51, 13150–13156.
33. N. Ballav, A. Maity and S. B. Mishra, Chem. Eng. J., 2012, 198–199, 536–546.
34. N. Ballav, H. J. Choi, S. B. Mishra and A. Maity, Appl. Clay Sci., 2014, 102, 60–70.
35. M. R. Mahmoudian, Y. Alias, W. J. Basirun and M. Ebadi, Appl. Surf. Sci., 2013, 268, 302–311.
36. H. Paudyal, B. Pangeni, K. Inoue, H. Kawakita, K. Ohto, K. Nath, H. Harada and S. Alam, Chem. Eng. J., 2013, 223, 844–853.
37. K. Biswas, D. Bandhoyapadhyay and U. C. Ghosh, Adsorption, 2007, 13, 83–94.
38. S. Deng, Z. Li, J. Huang and G. Yu, J. Hazard. Mater., 2010, 179, 1014–1021.
39. A. Dhillon, M. Nair, S. K. Bhargava and D. Kumar, J. Colloid Interface Sci., 2015, 457, 289–297.
40. G. Zhang, Z. He and W. Xu, Chem. Eng. J., 2012, 183, 315–324.
41. M. Karthikeyan, K. K. Satheeshkumar and K. P. Elango, J. Hazard. Mater., 2009, 163, 1026–1032.
42. R. Liu, W. Gong, H. Lan, Y. Gao, H. Liu and J. Qu, Chem. Eng. J., 2011, 175, 144–149.
43. M. Arshadi, Chem. Eng. J., 2015, 259, 170–182.
44. S. Azizian, J. Colloid Interface Sci., 2004, 276, 47–52.
45. N. Chen, Z. Zhang, C. Feng, N. Sugiura, M. Li and R. Chen, J. Colloid Interface Sci., 2010, 348, 579–584.
46. W. H. Cheung, Y. S. Szeto and G. McKay, Bioresour. Technol., 2007, 98, 2897–2904.
47. A. Ghosh, S. Chakrabarti, K. Biswas and U. C. Ghosh, Appl. Surf. Sci., 2014, 307, 665–676.
48. K. Babaeivelni and A. P. Khodadoust, J. Colloid Interface Sci., 2013, 394, 419–427.
49. Z. Wan, W. Chen, C. Liu, Y. Liu and C. Dong, J. Colloid Interface Sci., 2015, 443, 115–124.
50. L. Chai, Y. Wang, N. Zhao, W. Yang and X. You, Water Res., 2013, 47, 4040–4049.
51. A. Teimouri, S. Ghanavati Nasab, S. Habibollahi, M. Fazel-Najafabadi and A. N. Chermahini, RSC Adv., 2015, 5, 6771–6781.
52. Z. Qiusheng, L. Xiaoyan, Q. Jin, W. Jing and L. Xuegang, RSC Adv., 2015, 5, 2100–2112.
53. K. Li, Y. Wang, M. Huang, H. Yan, H. Yang, S. Xiao and A. Li, J. Colloid Interface Sci., 2015, 455, 261–270.
54. K. Y. Foo and B. H. Hameed, Chem. Eng. J., 2010, 156, 2–10.
55. A. L. Srivastav, P. K. Singh, V. Srivastava and Y. C. Sharma, J. Hazard. Mater., 2013, 263, 342–352.
56. H. Cai, G. Chen, C. Peng, Z. Zhang, Y. Dong, G. Shang, X. Zhu, H. Gao and X. Wan, Appl. Surf. Sci., 2015, 328, 34–44.
57. P. Saha and S. Chowdhury, Insight Into Adsorption Thermodynamics, Thermodynamics, ed. M. Tadashi, ISBN: 978-953-307-544-0, In Tech, 2011, pp: 349–364, http://www.intechopen.com/books/thermodynamics/insight-into-adsorption-thermodynamics.
58. V. J. Inglezakis and A. A. Zorpas, Desalin. Water Treat., 2012, 39, 149–157.
59. M. Mohapatra, D. Hariprasad, L. Mohapatra, S. Anand and B. K. Mishra, Appl. Surf. Sci., 2012, 258, 4228–4236.
60. Y. Cantu, A. Remes, A. Reyna, D. Martinez, J. Villarreal, H. Ramos, S. Trevino, C. Tamez, A. Martinez, T. Eubanks and J. G. Parsons, Chem. Eng. J., 2014, 254, 374–383.
61. M. Barathi, A. Santhana Krishna Kumar, C. U. Kumar and N. Rajesh, RSC Adv., 2014, 4, 53711–53721.
62. N. Thakur, S. A. Kumar, H. Parab, A. K. Pandey, P. Bhatt, S. D. Kumar and A. V. R. Reddy, RSC Adv., 2014, 4, 10350.
63. L.-X. Li, D. Xu, X.-Q. Li, W.-C. Liu and Y. Jia, New J. Chem., 2014, 38, 5445–5452.
64. X. Dou, Y. Zhang, H. Wang, T. Wang and Y. Wang, Water Res., 2011, 45, 3571–3578.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra20151b

---