Adsorption of phosphate on iron oxide doped halloysite nanotubes

Excess phosphate in water is known to cause eutrophication, and its removal is imperative. Nanoclay minerals are widely used in environmental remediation due to their low-cost, adequate availability, environmental compatibility, and adsorption efficiency. However, the removal of anions with nanoclays is not very effective because of electrostatic repulsion from clay surfaces with a net negative charge. Among clay minerals, halloysite nanotubes (HNTs) possess a negatively charged exterior and a positively charged inner lumen. This provides an increased affinity for anion removal. In this study, HNTs are modified with nano-scale iron oxide (Fe2O3) to enhance the adsorption capacity of the nanosorbent. This modification allowed for effective distribution of these oxide surfaces, which are known to sorb phosphate via ligand exchange and by forming inner-sphere complexes. A detailed characterization of the raw and (Fe2O3) modified HNTs (Fe-HNT) is conducted. Influences of Fe2O3 loading, adsorbent dosage, contact time, pH, initial phosphate concentration, and coexisting ions on the phosphate adsorption capacity are studied. Results demonstrate that adsorption on Fe-HNT is pH-dependent with fast initial adsorption kinetics. The underlying mechanism is identified as a combination of electrostatic attraction, ligand exchange, and Lewis acid-base interactions. The nanomaterial provides promising results for its application in water/wastewater treatment.

sorbent regeneration aspect when these are spent; which is a unique advantage over the other more commonly used materials, e.g., activated carbon 17 .
Halloysite nanotube (HNT) is a type of naturally occurring aluminosilicate with a nano-tubular structure that has a negatively charged exterior and a positively charged internal lumen. The HNT chemical structure is similar to kaolin (i.e., Al 2 Si 2 O 5 (OH) 4 ·nH 2 O), and is comprised of octahedral alumina crystals in the inner layer and a tetrahedral silica in the outer layer 16 . Its attractive features include high specific surface area, availability of micropores, presence of positive and negatively charged surface sites, and low-cost 18 . These attributes allow for HNTs to be integrated into various applications; e.g., HNTs are used as reinforcement fillers in polymers 19 , as drug delivery agents 20,21 , and as nanoreactors/nanotemplates for the synthesis of functional materials 22 as well as for applications in environmental remediation 23,24 . The use of HNT as a viable sorbent has recently gained interest, and these tubular minerals have been shown to remove various heavy metals from water 23 .
There is a paucity of studies on removal of anionic phosphate species with HNT. The unique surface properties (i.e., the positively charged inner lumen) of HNTs render it to be a promising material for electrostatic removal of phosphates and other anions. The exterior of HNT is considered to be chemically inert, except for the hydroxyl functional groups on the sides and on the broken edges. These exterior negative functional moieties can present electrostatic repulsion to anions, and thus can decrease the efficacy of phosphate sorption. Shielding such negatively charged sites with a highly absorptive oxide surface, such as hematite, can be advantageous.
Hematite, a major component of fine-grained sediments, is known to have a strong affinity for phosphate anions. Chemical modification with hematite can increase the number of active sites on HNT exteriors, improve the ligand exchange properties, and may present new functional groups that favor phosphate adsorption 25,26 . Specifically, the ligand exchange process allows for adsorption of phosphate onto hematite surfaces. Relatively strong inner-sphere complexes form with hematite surface sites and phosphate, and allow for specific interactions between the sorbate and the sorbent sites [27][28][29][30] . The study design thus considers distributing hematite surfaces by an in-situ growth of these oxides onto nano-scale HNT exterior. This way, aggregation of hematite (with no other surface modification) can be avoided, which can lead to an improvement in adsorption capacity (for phosphate) of these tubular minerals 25,31,32 . Hematite, alongside with clay, is also a low-cost and environmentally benign material and thus is an attractive choice for a sorbent 33 . A synergistic removal of phosphate by a positively charged HNT interior and iron oxide modified exterior could prove to be a promising low-cost nano-sorbent alternative.
This study presents a facile synthesis of Fe 2 O 3 -HNTs with benign precursors and in mild reaction conditions. No previous study to-date has reported surface modification of HNTs with ferric oxides. The nanomaterials are characterized with scanning electron and scanning transmission electron microscopy (SEM and STEM), transmission electron microscopy (TEM), BET surface area analysis, X-ray fluorescence (XRF), X-ray diffraction (XRD), and surface charge analysis. The application of raw and hybridized HNTs for phosphate removal from water is reported. The effects of sorbent dosage, contact time, pH, initial phosphate concentration, and coexisting ions have been systematically studied. Well-known sorption kinetics and equilibrium sorption models are used to assess the underlying mechanisms of adsorption.

Experimental Section
Materials. All solutions were prepared with reagent grade chemicals and deionized water (Milli-Q system).

Preparation of modified HNT (Fe-HNT).
A modified sol-gel method was used to surface modify HNTs.
In brief, 3 g of HNT was dispersed in 300 mL Milli-Q water and was magnetically stirred for 1 h. A desired amount (corresponding to the percent iron loadings) of FeCl 3 ·6H 2 O was dissolved in 150 mL Milli-Q water. A diluted ammonium solution was added to the iron solutions so that the moles of hydroxides (from NH 4 OH) are three times to the moles of iron in the solution and could achieve appropriate ratio to form hydroxyiron (Fe(OH) 3 ). The hydroxyiron solution was then added drop-wise to the HNT mixture and mixed at 350 rpm for 24 h. The mixture was then separated by centrifugation and washed with Milli-Q water several times prior to drying in air, overnight. The dried sample was collected and exposed to glacial acetic acid vapors in a furnace at 80 °C for 2 h. After exposure, the sample was left to dry at the same temperature for 30 min to remove any surface-sorbed acetic acid. The sample was then calcinated at 400 °C for 1 h to obtain the Fe-HNT. Finally, the cooled sample was sieved (100 μm sieve) to remove any sintered or agglomerated fraction. The Fe-HNT samples, prepared at different iron loadings of 0.25, 0.5, 1, and 5 wt.%, are designated in the text as 0.25Fe-HNT, 0.5Fe-HNT, 1Fe-HNT, and 5Fe-HNT, respectively.
Characterization. The surface morphology of the samples was studied with a JEOL JSM-7610F field emission SEM at an accelerating voltage of 5 kV. TEM images were obtained by placing the sample on lacey carbon film using an FEI Talos F200X TEM, and operating the TEM at 200 kV; the TEM is equipped with an STEM and an energy dispersive X-ray spectroscope (EDX). The specific surface area of HNT and modified HNT was meas- Aqueous Phosphate Analysis. Aqueous samples were analyzed for phosphate with an ion chromatography (IC) unit (Dionex ICS5000). Standards with known phosphate concentration were analyzed to obtain a www.nature.com/scientificreports www.nature.com/scientificreports/ calibration curve to ensure accuracy in analysis. At the end of the adsorption experiments, the samples were filtered using a 0.22 µm filter and a small aliquot was diluted with DI water at a desired concentration range (between 1-100 ppm) for analysis with IC.
Batch Adsorption Protocol. Unless stated otherwise, 3 g L −1 of the adsorbent was placed in a centrifuge tube with a 10 mg L −1 phosphate solution. The pH of the solution was adjusted with 0.1-1 mg L −1 HCl or NaOH. All samples were mechanically mixed on a shaker table (at 350 rpm in HBKU, Qatar and at 240 rpm in Austin, TX) at room temperature. For the kinetics and adsorbent dosage studies, the pH of the initial phosphate solution (i.e., pH of 5.0) was not altered (to avoid external chemical perturbation). The most efficient sorbent dose was determined from studies performed with sorbent amount ranging between 0.1 to 8.0 g L −1 . The sorbent dose used for the kinetics and equilibrium experiments was 3 g L −1 as this was found to be the most efficient amount for phosphate removal (Fig. S1). Kinetic experiments were conducted at time intervals ranging between 0.5 to 240 min to determine the equilibrium contact time and maximum adsorption capacity. Experiments investigating the effect of pH on adsorption capacity were conducted at a pH range of 2.0 to 10.0. The effect of initial phosphate concentration was determined at initial phosphate concentrations ranging between 0.5 mg L −1 to 100 mg L −1 and at a fixed pH of 4.0.
The adsorption capacity, q t , at a specific time t and the percent removal of phosphate were calculated based on the following equations: here, C 0 (mg L −1 ), and C t (mg L −1 ) denote the initial and equilibrium phosphate concentrations, respectively, V (L) is the volume of the solution, and W (g) is the mass of the adsorbent used.
Kinetics and Equilibrium Models. In order to evaluate the maximum phosphate uptake and identify the potential rate-controlling steps, three kinetic models (pseudo-first order, pseudo-second order, and intra-particle diffusion models) were applied to capture the adsorption process (on both HNT and 1Fe-HNT). The pseudo first order and pseudo second-order kinetics models are commonly used to obtain information on the equilibrium adsorption capacity of adsorbents. The model that provides the best fit and correlation coefficient is usually used to determine the adsorption capacity 34 . The details of these kinetic models are provided in the Supplementary Information.
The pseudo first order and pseudo second order kinetics models provide limited insights into the diffusion mechanism underlying this adsorption process. The Weber and Morris intraparticle diffusion model was used to identify the steps that occurred during the adsorption process and to elucidate whether intra-particle diffusion is the rate-limiting factor. The experimental data was therefore further tested against the Weber and Morris intra-particle diffusion model 35 which can be expressed as follows: where, q t (mg.g −1 ) is the amount of phosphate adsorbed at time t (min), k p is the intraparticle diffusion rate constant (mg.g.min 0.5 ) and C is a constant. The values of k p and C can be determined from the intercept and slope of the linear plot of q t versus t 0.5 . In order to further understand the adsorption mechanism, classical adsorption isotherm models (i.e., Langmuir and Freundlich) were applied to fit the experimental data. These equilibrium models highlight sorbate-sorbent binding interaction and also give insights into possible mechanisms of adsorption. The Langmuir isotherm is based on monolayer adsorption onto the active sites 36 . An important characteristic of the Langmuir model is the dimensionless constant (R L ), generally known as the separation factor, which was also estimated. The value of R L indicates whether adsorption is irreversible (R L = 0), favorable (0 < R L < 1), linear (R L = 1), or unfavorable (R L > 1). The Freundlich isotherm, on the other hand, captures a non-ideal and reversible adsorption process not restricted to monolayer adsorption. This empirical model assumes a heterogeneous surface and that the amount adsorbed increases with sorbate concentration 37 . The model and its parameter details are outlined in the Supplementary Information. For maintaining consistency, all phosphate in the text is represented as orthophosphate (PO 4 ).

Results and Discussion
Physical Morphology. SEM images of raw HNT and 1Fe-HNT (Fig. S2a,b) reveal the tubular structure of these clays. Clearer illustrations of the structure of the raw HNT and 1Fe-HNT are obtained through HRTEM imaging, as shown in Fig. 1a,b, respectively. 1Fe-HNT is chosen for TEM characterization since it is used for the equilibrium experiments. TEM images of the raw HNTs (Fig. 1a) illustrate the tubular structures of nanoclay, with open ends and a hollow cavity (lumen). A significant difference in HNT size, before and after modification (Fig. 1b), is not observed. For the 1Fe-HNTs shown in Fig. 1b, it is observed that the Fe 2 O 3 nanoparticles are attached on the surface of the HNTs. The average diameter of the Fe 2 O 3 nanoparticles on the surface of HNTs is determined to be 5.6 ± 0.92 nm.
STEM micrographs of 1Fe-HNT (Fig. 1c,d) confirm the presence of Fe, which is distributed throughout the surface of the HNTs. Since the surfaces of HNTs are negatively charged aluminosilicates, these could serve as www.nature.com/scientificreports www.nature.com/scientificreports/ polyanionic surfaces to allow for complex formation with iron cations 38 . The length of the raw HNTs ranges between 68 to 1520 nm (Fig. 1e), while the external diameter ranges between 20 to 150 nm (Fig. 1f). The lumen diameter is found to vary between 10 and 28 nm (Fig. 1g).
Relative to raw HNT, the specific surface area of the iron oxide modified HNTs is found to increase slightly from 64.4 to 70.5 m 2 /g (Table S1). This slight increase could be attributed to the contribution of nano-sized iron oxide particles, hybridized onto the HNTs. Figure S3 presents the N 2 adsorption/desorption isotherms of HNT and modified HNTs at different iron oxide loadings. According to the classifications of International Union of Pure and Applied Chemistry (IUPAC), all isotherms of the raw and modified HNTs were of type II 39 with H3 hysteresis loops 40 . The type II isotherm is indicative of a macroporous structure, however the hysteresis loop of the type H3 ascribes materials that have slit-shaped pores 41 .
Chemical Composition and Crystallinity. Table S2 presents the chemical composition of the HNTs, evaluated with X-ray fluorescence. HNTs are primarily composed of silicon dioxide and aluminum dioxide with a trace amount of CaO, SrO, TiO 2 , phosphorous pentoxide, and sulfur trioxides. Raw HNTs contain a certain amount of Fe 2 O 3 (0.59 wt.%), which increases with the increase in reagent loading. Iron content is also shown to increase upon hybridization. The amount of iron loading for 0.2, 0.5, 1, and 5Fe-HNT can be deduced from the table to be 0.92, 1.85, 2.47, and 5.98 wt.%, respectively.
The XRD patterns of raw HNT and iron oxide modified HNTs are shown in Fig. 2. Clay minerals are primarily distinguished by the noticeable basal cleavage of the layered silicate structures. The first order basal reflection (001) for raw HNT is at 7.49Å, which is indicative of a kaolin-type mineral and of dehydrated halloysite 42 44,45 . The presence of these peaks confirms the presence of iron oxides on the raw and modified HNTs.
Surface Charge Density and Effect of pH on Adsorption Behavior. The electrophoretic mobility of HNT and 1Fe-HNT as a function of solution pH is shown in Fig. 3a. The positive charge at a low pH for raw HNT can arise from protonation of the hydroxyl groups on the clay edges or at surface defects sites 46 . The point of zero charge (PZC) for raw HNT is determined to be 2.5, which is consistent to the values reported earlier 47,48 .
An interesting feature of HNT should be noted; i.e., the chemical composition difference between the exterior and interior of the HNTs, dominated by SiO 2 (negatively charged) and Al 2 O 3 (positively charged), respectively. It has been reported that the inner lumen consisting of Al 2 O 3 maintains a positive charge throughout the pH range of 2.5 to 8.5 49 . This property allows for the selective adsorption of anions in the lumen in the pH range studied.
After modification of the HNTs with iron oxides, the PZC shifted to 3.3. Modifying HNT with iron oxide induces a shift towards a more positive value in mobility in most of the pH values studied, which is consistent to www.nature.com/scientificreports www.nature.com/scientificreports/ the previous observations with hematite-modified clay 50 . It is worth noting that the PZC of hematite is 5.8 as published earlier, suggesting that the hematite surfaces hold a positive charge over a wider pH range 50,51 . This phenomenon was also observed by Arias et al. 51 , and was attributed to the adsorption of hematite on the negatively charged silica basal surface of kaolinite resulting in the reduction of the number of negative charges on kaolinite 51,52 . The surface potential determination suggests that the reduction in the magnitude of the negative charge of 1Fe-HNT has likely occurred from the iron oxides bound to the HNT exterior; mutual charge neutralization probably have resulted between oppositely charged surface moieties (in HNTs and iron oxide) 50,51,53 . Another theory could be that the negative charges on the HNT are physically blocked by the iron oxides upon hybridization, indicating successful attachment of the iron oxides onto the HNT surfaces; a previous study involving hematite-kaolinite complexes reported a similar observation 54 .
Since iron dissolution at low pH can be a concern, dissolved iron content is analyzed to rule out such potential complication. Dissolution of SiO 2 has predominantly been reported to take place at pH < 2.0 and at pH > 9.0 55 . Now, AlO 2 is known to be relatively insoluble in acids and in strong alkali solutions 56,57 . Also, previous reports, investigating the effects of pH changes on hematite, conclude that the solubility of hematite is extremely low at pH > 3.0 58 . Experiments conducted at a pH of 2.0 with 1Fe-HNT and at a contact time of 240 min revealed the final iron concentration in the solution to be less than 0.05 mg/L, when filtered with a 0.45 µm filter; this indicates that the composite material maintains its integrity at a low pH.
The adsorption of phosphate onto raw HNT and 1Fe-HNT has been studied over a pH range of 2.0 to 10.0 (Fig. 3b). For raw HNT, the adsorption increased with an increase in pH up to 4.0, beyond which, the change in adsorption is insignificant. At pH 2.0, i.e., below the PZC of HNT, the dominant phosphate species are H 3 PO 4  , and are present in nearly equal concentrations. The adsorption of phosphate at these conditions can be attributed to electrostatic attraction between the positively charged sorbent exterior and negatively charged phosphate species. The low adsorption capacity could be attributed to a strong presence of uncharged H 3 PO 4 species. However, at pH > PZC, the external surface of raw HNT is negatively charged (upon deprotonation) and the adsorption capacity is found to be 1.5 mg.g −1 . This indicates that phosphate is likely been removed via electrostatic attraction between the anionic species and the positively charged inner lumen of HNTs.
The change in pH affects the adsorption capacity of 1Fe-HNT sorbents in a unique way. It is known that hematite modification of kaolinite can induce a change in both its chemical and physical properties 27 . Adsorption of phosphate is observed to have occurred over the entire pH range (i.e., 2.0 through 10.0) for both HNT and 1Fe-HNT. HNTs possess a significant amount of hydroxyl groups on the edges 16 . Modification of HNTs with iron oxide nanoparticles increase the hydroxyl groups available on the surface and edges and thus can lead to an increase in the adsorption capacity of the sorbent. At pH 4.0, the dominant phosphate species is the mono-ionic H 2 PO 4 − (98%). From Fig. 4 it appears that 1Fe-HNT and raw HNT both have an affinity for phosphate adsorption for this species. The underlying mechanism of such decrease in phosphate sorption at higher pH may be explained by the electrostatic repulsion between the negatively charged HNT or 1Fe-HNT exterior and the phosphate anions. Such repulsion has likely prevented interaction of the phosphate anions with the inner lumen of the clay, resulting in a decreased capacity. Furthermore, competition for adsorption sites between the OHions and phosphate species may occur at an alkaline pH range 59 .
The Kinetics of Adsorption. A significantly rapid adsorption is observed within the first 30 s (see inset of Fig. 4a) and equilibrium is reached in 4 h. The rapid adsorption in the initial stage can be attributed to the unique structure of the HNTs. As shown in the electron micrographs, HNTs have large pore diameters ranging between 10 to 30 nm, which can allow easy access of phosphate anions and binding in the inner lumen.
The kinetic data fit with pseudo-first order and pseudo-second order models are shown in Fig. 4b,c and the key parameters are listed in Table 1. It is known that the pseudo-first order model provides a better fit for the initial stage of the reaction process, specifically for those sorbents with rapid adsorption ability 60 . The correlation coefficient (R 2 ) value for the pseudo-first order model is relatively lower than that for the pseudo-second order model. Furthermore, the calculated value of q e,calc (mg.g −1 ) for 1Fe-HNT is significantly lower than the experimental value of q e , exp (mg.g −1 ), obtained from the pseudo-first order model. This indicates that the pseudo-first order kinetic model is not suitable to capture the adsorption process.
Conversely, the calculated equilibrium capacities for HNT and 1Fe-HNT obtained from the pseudo-second order model 61 (Fig. 4b) are very similar to those estimated from the experimental analysis (as shown in Table 1). In addition, the higher correlation coefficients of raw HNT (R 2 = 0.995) and 1Fe-HNT (R 2 = 0.997) further suggest that the adsorption of phosphate onto HNTs and iron oxide modified HNTs follows the pseudo-second order A plot of q t versus t 0.5 yielding a linear relationship that passes through the origin indicates that intraparticle diffusion is the sole rate-limiting step in the reaction 35 . As shown in Fig. 4c, the linear plots for both HNT and 1Fe-HNT deviate from the origin; which indicate that intraparticle diffusion is involved in the adsorption process but is not the only rate-determining step. The value of C provides an insight onto the thickness of the boundary layer. In general, the larger the C value, the greater is the boundary layer effect (i.e., the greater is the contribution of surface adsorption in the rate-determining step), while a value of zero indicates that intra-particle diffusion dominates throughout the adsorption process. Also, higher values of C typically indicates higher adsorption 63 .
The q t versus t 0.5 plot of raw HNT shows the existence of a single rate while the plot for 1Fe-HNT shows two linear regimes; which implies a multi-stage adsorption process 14 . The initial linear regime is a result of external surface adsorption, driven by the initial phosphate concentration. The second linear regime signifies that phosphate species diffusing within the pores of the hollow HNT over time; which is typical in the case for intraparticle diffusion. This result is significant in that it provides further evidence for enhanced adsorption of phosphate on the exterior of the iron-oxide modified HNTs as opposed to that within the interior lumen (in case of raw HNTs). Table 1 shows that the rate constant for 1Fe-HNT in the first regime (k p1 ) is greater than the second (k p2 ); this indicates that intraparticle diffusion is the rate determining step of the entire adsorption process 64 .
The Sorption Equilibrium Study. The Langmuir and Freundlich isotherms are presented in Fig. 5 and the key parameters are listed in Table 2. The highest correlation coefficients for HNT (R 2 = 0.973) and 1Fe-HNT (R 2 = 0.955) are derived by fitting the equilibrium data with a Langmuir isotherm model. The monolayer capacities are near the experimental adsorption capacities, indicating an agreement with the Langmuir model. The calculated values for R L of 0.070 and 0.029 for HNT and 1Fe-HNT, respectively; which confirm that the adsorption process is favorable 65 . These results indicate that the surfaces of HNTs and 1Fe-HNTs are relatively uniform and that a monolayer of phosphate coverage dominates the adsorption process. It has been observed in previous studies 66 that anion adsorption at any pH value increases with the increase in sorbate concentration. The Freundlich isotherm model showed lower correlation coefficient values but R 2 values were greater than 0.90 for both sorbents ( Table 2). The Freundlich constant K F is associated with the adsorption capacity of the sorbent; a higher value indicates a higher affinity for the sorbate 64 .
Effect of Coexisting Anions. In water and wastewater, anions such as sulfate, nitrate, and chloride coexist with phosphate ions and will likely compete for adsorption sites. The effect of these anions on the adsorption capacity is investigated and the results are shown in Fig. 6. It is observed that the addition of the coexisting anions in the solution does not impede the removal of phosphate from water, rather slightly enhance it.
Although the presence of anions in a solution has generally been reported to compete for adsorption sites, the total adsorption capacity of some metal oxides has been found to be unaffected or has increased in some cases 25,67 . Specific additives, such as NaCl, NaNO 3 , and KNO 3 have been found to enhance phosphate removal 25 . It can also be inferred that chloride may contribute in the adsorption of phosphate from the solution 25 . A similar phenomenon was reported by Zhang and coworkers 68 and Giesler et al. 29 who used other metal oxides for phosphate adsorption. It is thus deduced that anions that adsorb via outer-sphere complexation, are highly sensitive to ionic strength. Consequently, the adsorption of these anions is inhibited by competition with other weakly www.nature.com/scientificreports www.nature.com/scientificreports/   To answer the question of why the sorbent with a negative surface charge has an affinity for anion sorption, it is important to address the concepts of both specific and non-specific adsorption. The results from the pH studies indicate an occurrence of specific adsorption (ligand exchange or chemisorption); i.e., formation of a strong covalent bond due to the attachment of the ion directly to the surface without a water molecule interposed in between 71 . Furthermore, the chemical component of the adsorption free energy dominates in specific adsorption; as a result, sorption occurs on positive, negative, or neutral surfaces 66,71 . On the other hand, when non-specific adsorption (i.e. electrostatic attraction) governs the adsorption process, the sorbent surface must have a positive overall charge in order for phosphate sorption to occur; consequently, adsorption can only occur at a pH lower than the PZC 66 , which is not the case in this study. Furthermore, at the low pH conditions, an increase in the final pH value is observed (Table S3). This indicates that phosphate adsorption is likely caused by exchange of ions; i.e., an exchange between the phosphate and the hydroxide ligands on the surface occurred, which released hydroxides into the solution. Also, ligand exchange can occur in conjunction with electrostatic attraction between the negatively charged phosphate anion and positively charged surfaces (for pH < PZC) and the inner lumen 27,66 . However, at an alkaline pH, the observed decrease in equilibrium pH can result from the deprotonation of the coordinated water molecules on the active site. Adsorption can occur via Lewis acid-base interaction in which the iron active site becomes a weak base (Lewis base) and the phosphate anions act as a weak acid (Lewis acid) 8,72 . Furthermore, during this interaction, a reactive iron site can react with oxygen in the phosphate anions to develop Fe-O coordination bonds, thus releasing an H 2 O molecule in the process as shown below. This mechanism was also proposed in earlier studies with iron(hydr)oxide sorbents 73 . Figure 7 provides a graphic illustrating the mechanisms of phosphate adsorption onto Fe-HNTs.
The complexation mechanism to describe the reaction of phosphate ions on raw HNT and iron oxide modified HNT is adopted from an earlier study 27 (5) and (6)  where, SOH denotes the surface functional group of halloysite and iron oxide modified halloysite. Table 3 lists a comparison of equilibrium and maximum adsorption capacities of Fe-HNT and other adsorbents that are considered low-cost, naturally available, or that exhibit similar chemical features. In general, the adsorption capacity of Fe-HNT is similar to or exceeds that of other low-cost adsorbents, such as iron hydr(oxide) coated sand, magnetic iron oxide nanoparticles, and other clays [74][75][76] . Although the achieved adsorption capacity in this study is lower than that of modified activated carbon, HNTs have a much lower cost compared to granular activated carbon and commercial activated carbon (Table 4). Sand, on the other hand, is low in cost, however, its adsorption capacity towards phosphate is poor. The time to reach adsorption equilibrium is an important factor, which can dictate the feasibility of an adsorbent for a treatment process; Fe-HNT shows a relatively short equilibrium time, relative to the other adsorbents.

Sorption Performance Comparison with Other Sorbents.
Considering the low cost and comparable adsorption capacity of Fe-HNT, this sorbent can be considered as a promising material for phosphate removal from water. Furthermore, its natural and environmentally friendly features make it suitable to be used as a fertilizer or soil conditioner. Selling the phosphate saturated adsorbent material as a fertilizer could be more cost effective than regenerating it; since the cost of regeneration usually account for more than 70% of the total operating and maintenance cost of an adsorption system 77 . However, future work should involve the study of regenerating the material with NaOH and precipitating and retrieving the phosphate from the solution using Ca(OH) 2 .

Conclusions
In this study, a facile preparation of a natural and environmentally benign iron oxide modified halloysite nanotubular clay (1Fe-HNT) is presented. TEM micrographs, diffractograms, and XRF spectrographs confirm successful hybridization of nano-scale iron oxide (α-Fe 2 O 3 ) onto HNT surfaces. Relative to raw HNT, the as-prepared nanosorbent displays an enhanced adsorption capacity (for phosphate). The adsorption capacity is found to be pH-dependent. Kinetics results reveal that the prepared sorbent can remove more than 30% of the phosphate within the first 30 s. The kinetics behavior closely follows a pseudo-second order process, and intraparticle diffusion is likely the rate-limiting step. A Langmuir model best describes the adsorption isotherm data. The presence of coexisting ions slightly increase the adsorption capacity of phosphate for 1Fe-HNT, specifically when chloride is present. This indicates strong selectivity of the sorbents towards phosphates. It can be deduced that electrostatic attraction and ligand exchange govern sorption at low pH, while Lewis acid-base interactions dominate the process as the pH increases. The results presented in this study demonstrate that 1Fe-HNT is a promising sorbent   Table 4. Cost of some of the compared sorbents.