FUNCTIONALIZED HALLOYSITE NANOTUBES FOR ENHANCED REMOVAL OF Hg2+ IONS FROM AQUEOUS SOLUTIONS

Water is essential for humans, animals, and plants; pollutants, usually derived from anthropogenic activities, can have a serious effect on its quality. Heavy metals are significant pollutants and are often highly toxic to living organisms, even at very low concentrations. Among the numerous removal techniques proposed, adsorption onto suitable adsorbent materials is considered to be one of the most promising. The objective of the current study was to determine the effectiveness of halloysite nanotubes (HNT) functionalized with organic amino or thiol groups as adsorbent materials to decontaminate polluted waters, using the removal of Hg2+ ions, one of the most dangerous heavy metals, as the test case. The effects of pH, ionic strength (I), and temperature of the metal ion solution on the adsorption ability and affinity of both materials were evaluated. To this end, adsorption experiments were carried out with no ionic medium and in NaNO3 and NaCl at I = 0.1 mol L−1, in the pH range 3–5 and in the temperature range 283.15–313.15 K. Kinetic and thermodynamic aspects of adsorption were considered by measuring the metal ion concentrations in aqueous solution. Various equations were used to fit experimental data, and the results obtained were explained on the basis of both the adsorbent’s characterization and the Hg2+ speciation under the given experimental conditions. Thiol and amino groups enhanced the adsorption capability of halloysite for Hg2+ ions in the pH range 3–5. The pH, the ionic medium, and the ionic strength of aqueous solution all play an important role in the adsorption process. A physical adsorption mechanism enhanced by ion exchange is proposed for both functionalized materials.


INTRODUCTION
Among toxic metals, Hg is considered to be one of the most dangerous to humans and all other living organisms as shown by the results of numerous toxicological studies (e.g. Bernhoft 2012; Ynalvez et al. 2016). For this reason, many research programs exist which are devoted to the monitoring and assessment of Hg diffusion in the environment, supported by national and international organizations (UN Environment 2019). Three types of Hg sources are responsible for environmental contamination, namely, natural, anthropogenic, and a third source related to the remobilization of previously settled Hg from soils and sediments (Wang et al. 2004). Based on a rough estimate, 30% of Hg emissions in the atmosphere are of anthropogenic origin and the amount of anthropogenic Hg released into natural waters is ~1000 tons per year (UN Environment 2019). Due to biogeochemical transformations, Hg in natural waters can be present in elemental (Hg 0 ), inorganic (Hg + , Hg 2+ ), or organic [CH 3 Hg + , (CH 3 ) 2 Hg] forms. All species of Hg react with other components of aquatic systems to complicate its speciation picture (Von Burg and Greenwood 1991). Among the various forms, methylmercury is considered to be the most toxic whilst the inorganic species, especially Hg 2+ , are the most soluble and abundant in polluted waters (Wang et al. 2004) Considering the persistence of the element in the environment, the only solution to the 'mercury problem' is the reduction of anthropogenic emission. Much work is being done to find a good remediation method for removal of Hg from wastewaters before its release into the environment. One of the most promising techniques for removing toxic metal ions from contaminated aqueous systems is adsorption onto materials of natural origin that are cheap, abundant, and non-toxic De Gisi et al. 2016). Among natural adsorbent materials, the unique features of clay minerals have attracted great interest. They are cheap, available in large quantities, have little or no toxicity of their own, and have a very small environmental impact (Dong et al. 2012;Cataldo et al. 2015). Several types of clay minerals have already been used as adsorbents of pollutants, natural or modified, in remediation studies and, among them, halloysite is one of the most interesting (Dong et al. 2012;Cataldo et al. 2015;2018;Peng et al. 2015;Renu et al. 2017;Uddin 2017).
Halloysite is a clay mineral with a mainly hollow tubular structure which consists of 10-15 aluminosilicate layers with outer and inner surfaces consisting of siloxane, silanol, and aluminol groups (Joussien et al. 2005). Due to this particular structure, HNT are positively charged inside and negatively charged outside over a wide pH range (Bretti et al. 2016). The charge separation and the tubular structure enable this clay mineral to be used in many applications, e.g. as a drug carrier and in drug delivery, as a catalyst support, as a filler for hydrogels and polymers, as an adsorbent of pollutants, etc. (Abdullayev et al. 2010;Dong et al. 2012;Kamble et al. 2012;Wei et al. 2013;Owoseni et al. 2014;Cataldo et al. 2018;Massaro et al. 2020).
In a recent study, a commercial halloysite and its functionalized form with amino groups were used to good effect as adsorbents of Pb 2+ ions ). In the present study, the objective was to test the same pristine halloysite (p-Hly) and two functionalized forms with amino groups (Hly-NH 2 ) and thiol groups (Hly-SH) as adsorbents for the removal of Hg 2+ ions from aqueous solutions, with the hypothesis, as proven in various previous studies (e.g. Cataldo et al. 2013Cataldo et al. , 2018, that the adsorption capacity of a material depends not only on its affinity toward the pollutant, but also on the experimental conditions (pH, ionic medium, ionic strength, temperature) of the aqueous solution containing the pollutant to be removed and that this aspect is even more important when the pollutant is a metal ion such as Hg 2+ .
The Hg 2+ ion forms quite stable hydrolytic and chloride species in the pH range typical of natural waters and industrial wastewaters and with the anion commonly present in these water matrices (Baes and Mesmer 1976;Martell and Smith 1977;Martell et al. 2004;Crea et al. 2014). With this in mind, adsorption experiments were carried out for both functionalized clay materials, adjusting the experimental conditions of the solutions containing Hg 2+ ions. The results together with the characterization of the clay materials Massaro et al. 2019) were analyzed in order to: (1) establish the effect of halloysite functionalization, the kinetics and thermodynamics of Hg 2+ ion adsorption, and the best experimental conditions in terms of adsorption ability of both materials; and (2) evaluate the mechanism of adsorption.

Chemicals and Materials
Pristine halloysite (p-Hly) was a commercial product (Sigma, lot MKBQ8631V) and was used after washing with ultrapure water (ρ ≥ 18 MΩ cm) and drying in an oven at T = 383.15 K. 3-azido propyltrimethoxisilane was synthesized as reported elsewhere (Massaro et al. 2016); 3-mercapto propyltrimethoxysilane and all reagents needed for the synthesis were purchased from Sigma-Aldrich (Steinheim, Germany) and used without further purification.
Sodium nitrate and sodium chloride pure salts (Fluka) were used, after drying at 383.15 K for 2 h, to adjust the ionic strength of solutions. Nitric acids and sodium hydroxide used to adjust the pH of the metal ion solutions were prepared by diluting concentrated Fluka solutions. Hg 2+ ion solutions were prepared by weighing the Hg(NO 3 ) 2 (Sigma-Aldrich, Steinheim, Germany), analytical grade salt. Mercury standard solutions used for calibration curves were prepared by diluting a 1000 mg L −1 standard solution in 2% HNO 3 (CertiPUR, Merck, Darmstadt, Germany). All the solutions were prepared using fresh, CO 2 -free, ultrapure water (ρ ≥ 18 MΩ cm) and grade A glassware.

Synthesis of Hly-NH 2 and Hly-SH Nanomaterials
The adsorbent materials were obtained following procedures reported previously Massaro et al. 2019). In particular, the pristine halloysite was reacted with the appropriate silane (3-mercaptopropyltrimethoxysilane or 3-azidopropyltrimethoxysilane, respectively) to give the compounds Hly-SH and Hly-N 3 (Fig. 1).
The latter was then subjected to reduction under the Staudinger reaction conditions (triphenylphosphine, DMF, r.t.), to obtain, finally, the Hly-NH 2 material with significant organic moiety loading. The degrees of functionalization of halloysite, estimated using thermogravimetric analysis, were 0.40 mmol g -1 and 0.07 mmol g −1 for Hly-SH and Hly-NH 2 , respectively. The successful functionalization was also verified by Fourier-transform infrared (FTIR) spectroscopy (Fig. 2).
The FTIR spectrum of the thiol-functionalized Hly (Fig. 2c) showed the typical Hly vibration stretching bands, (Massaro et al. 2016) and, in addition, exhibited the vibration bands for C-H stretching of methylene groups around 2980 cm −1 attributed to the organic functionalities introduced. In the FTIR spectrum of Hly-NH 2 (Fig. 2a), in addition to the aforementioned vibration stretching bands, a strong vibration band at ~ 1500 cm −1 due to the bending vibrations of the -NH 2 groups was also present Massaro et al. 2018).

Experimental Equipment and Procedures for Kinetic and Thermodynamic Experiments
Batch experiments were carried out in order to study the kinetic and thermodynamic aspects of Hg 2+ adsorption onto p-Hly, Hly-NH 2 , and Hly-SH.
In kinetic experiments, ~15 mg of adsorbent material was placed in 12 Erlenmeyer flasks containing 25 mL of Hg(NO 3 ) 2 solution (c Hg 2+ ≈ 40 mg L −1 ), at pH = 4, with no added ionic medium, and at T = 298.15 K. The Hg 2+ ion concentration in solution was measured at various adsorbent-solution contact times over the time interval 0-360 min. The pH of solution was monitored during the experiments.
Isotherm experiments were carried out in the pH range 3-5, with no ionic medium, in 0.1 mol L −1 NaNO 3 , in NaCl (with Hly-SH adsorbent only), and over the temperature range 283.15-313.15 K. In each equilibrium experiment, ~15 mg of adsorbent material was placed in nine Erlenmeyer flasks containing 25 mL of Hg(NO 3 ) 2 solution at various concentrations (5 ≤ c Hg 2+ (mg L −1 ) ≤ 50). The solutions were stirred at 180 rpm for 12 h using an orbital mixer (model M201-OR, MPM Instruments, Bernareggio, Italy) and then were separated from the adsorbent before measuring the pH and the Hg 2+ concentration. The concentrations of Hg 2+ ions in the solutions collected in kinetic and isotherm experiments were measured by Inductively Coupled Plasma Optical Emission Spectroscopy (ICP-OES) using a Perkin Elmer (Waltham, USA) Model Optima 2100 instrument, equipped with an autosampler model AS-90. The Hg emission intensity was measured and each measurement was repeated three times. Calibration curves were measured in the same experimental conditions and covering the Hg 2+ ion concentration range of kinetic and thermodynamic adsorption experiments. The pH of the Hg 2+ ion solutions was measured with a combined ISE-H + glass electrode (Ross type 8102, Thermoscientific, Waltham, USA). The ISE-H + electrode was calibrated previously under the same experimental conditions as the adsorption experiments. To this end, 25 mL of standardized HNO 3 solution was titrated with NaOH using a potentiometric titration system (Metrohm, Herisau, Switzerland, Model 888 Titrando) controlled by the TIAMO software (version 2.3 light, Metrohm -Herisau, Switzerland).

Kinetic and Isotherm Models for Hg 2+ Ion-adsorption Studies
Kinetic data were fitted with the pseudo-first order equation (PFO) of Lagergren (Eq. 1), the pseudo-second order equation (PSO) (Eq. 2), and the intraparticle diffusion equation of Vermeulen (Ver) (Eq. 3) (Lagergren 1898;Blanchard et al. 1984;Guo et al. 2008): where q t and q e represent the adsorption capacity of the p-Hly, Hly-NH 2 , or Hly-SH (mg g −1 ) at time t and at equilibrium and k 1 (min −1 ), k 2 (g mg −1 min −1 ), and k v (min −1 ) are the rate constants of adsorption. In the present study, the integrated forms of equations for the boundary conditions t = 0 to t = t and q t = 0 and q t = q t , listed below, were used: The adsorption equilibrium data were processed with Freundlich (F) (Eq. 7) and Langmuir (L) (Eq. 8) isotherm equations (Freundlich 1906;Langmuir 1918): where q m (mg g −1 ) is the maximum adsorption ability of the adsorbent, c e (mg L −1 ) is the Hg 2+ concentration in solution at equilibrium; K F (L 1/n g −1 mg 1−1/n ) and K L (L·mg −1 ) are the constants in the Freundlich and Langmuir models, respectively. The Hg 2+ ion-adsorption capacity at various contact times t (q t , mg g −1 ) in kinetic experiments, or at different Hg 2+ / adsorbent ratios in the thermodynamic studies (q e , mg g −1 ) were calculated using Eq. 9: where V (L) is the volume of the Hg 2+ solution and m is the mass (g) of the adsorbent material (p-Hly, Hly-NH 2 , or Hly-SH); c 0 and c t are the Hg 2+ ion concentrations (mg L −1 ) at t = 0 and t = t, respectively. At equilibrium conditions, Eq. 9 was used by replacing c t with c e to calculate q e .
where R is the universal gas constant, 0.008314 kJ mol −1 K −1 , and T is the temperature in K.
The LIANA and OriginLab suite software (OriginLab Corporation, Northampton, Massachusetts, USA), were used to fit kinetic and isotherm equations to experimental data.

Speciation Analysis
The percentage of protonated/unprotonated functional groups of the adsorbent material, the charge of the species formed by the metal ions, and the experimental conditions (ionic medium, ionic strength, pH) for the aqueous solutions to be treated all play an important role in the efficiency of metal-ion sorption. Opposite charges for the metal ion species and for the surface of adsorbent material undoubtedly favor the adsorption process via Coulombic attraction. For this reason, knowledge of the species distribution of both metal ions and adsorbent material is required in order to discuss the adsorption results effectively.
According to previously reported studies on the acidbase properties of the pristine material, aluminol groups are fully protonated in the pH range investigated in this work, while the silanols are partially deprotonated: 50% of Si-OH groups are protonated at pH = log K H ≈ 4, with increasing fraction of protonated and deprotonated groups below and above this pH value, respectively (Bretti et al. 2016).
Other experiments need to be performed using pure water (no ionic medium added) and solutions containing Na + salts as the ionic medium (i.e. NaNO 3 and NaCl at I = 0.1 mol L -1 ). In fact, as already reported, Na + ions form weak ion pairs with silanol groups (Bretti et al. 2016). Despite their weakness, relatively large Na + ion concentrations (relative to that of the sorbent material) led to formation percentages of sodium ion pairs that are not negligible and, interestingly, modify the charge of material surface and reduce the number of available binding sites (occupied by sodium itself).
For the materials functionalized with -NH 2 or -SH groups, though the protonation constants have not been determined experimentally, both amino and thiol functions can be assumed to be fully protonated in the abovementioned pH range (3-5), as aliphatic amines and thiols generally show log K H >> 6 (e.g. log K H ≈ 10.6 and 10.9 for propylamine and propane 2-thiol, respectively (Martell et al. 1977(Martell et al. , 2004). Of course, complexation by metal cations may result in proton displacement from the protonated functional groups. The soft nature of S donors should facilitate H + exchange by Hg 2+ , although -NH 2 groups also have good affinity toward these cations. In fact, from a rough estimate of the contribution of the Sand N-donors to the logK HgL of a generic ligand L, based on the stability constant values of 143 amines and 92 thiols, on average, the contribution of -SH and -NH 2 groups to the stability of the HgL species has been shown to be 22.3 and 6.4 log units, respectively (Crea et al. 2014).
Another important aspect to take into account in evaluating the results obtained here is the speciation of Hg 2+ ions, which are subjected to a strong ionic medium effect due to the ability of Hg to form stable chloride complexes (Crea et al. 2014). In pure water or low-interacting media (e.g. NaNO 3 , see speciation diagram in Supplementary Material Fig. S1a), Hg is present as free cations and/or in its hydrolyzed forms (Hg(OH) + and Hg(OH) 2 ), in percentages which depend on pH. Hydrolysis is suppressed completely at pH < 6.0 in NaCl media, in which HgCl r (2−r) (with 1 ≤ r ≤ 4) complexes dominate Hg speciation, with ratios between these species depending on Cl − concentration (e.g. in NaCl at I = 0.1 mol L -1 , the percentages of chloride species are: HgCl 2 ≈ 38%, HgCl 3 − ≈ 44%, HgCl 4 2− ≈ 18%) (see speciation diagram in Fig. S1b) (Baes and Mesmer 1976;Crea et al. 2014).

Kinetics of Adsorption of Hg 2+ Ions onto Hly-NH 2 and Hly-SH Materials
The kinetics of the adsorption of Hg 2+ ions onto the two functionalized halloysite materials were studied in solution with no ionic medium, at initial pH = 4. The kinetic data have been fitted with PFO, PSO, and Ver kinetic models.
The kinetic constants and q e values are reported in Table 1 together with the experimental q e (measured after 24 h), and the corresponding R 2 and standard deviation values of the fits. From comparison of statistical parameter values, although all the models gave a quite good fit, the Ver and the PFO can be considered the best models (the largest R 2 and the smallest standard deviation values of the fits) for Hg 2+ adsorption onto Hly-SH and Hly-NH 2 adsorbents, respectively. The goodness of model fits were confirmed by comparisons between the experimental and calculated q e values. The experimental data together with the fitted curves of the three kinetic equations are reported in Fig. 3 for both of the Hg 2+ -adsorbent systems. Differences in adsorption kinetics of the two adsorbents can be attributed to their different degrees of functionalization (see the sections above on 'Synthesis of Hly-NH 2 ' and 'Hyl-SH Nanomaterials'). Because the thiols are soft ligands, they have a greater affinity for the metal ion as confirmed by the greater q e values. For both of the adsorbent materials, adsorption equilibrium is reached after ~200 min. The results obtained with Hly-SH agree with those found by Manohar et al. (2002) for the Hg 2+ ion adsorption onto a clay (Thonnakkal Clay Mine; >90% kaolinite and mica) functionalized with thiol groups with 2-mercaptobenzimidazole. The current authors found a q e of 23.839 mg g −1 and a k 2 of 2.011 10 -3 g mg −1 min −1 at pH = 6, T = 30°C, and c Hg2+ = 50 mg L −1 .

Modeling the Equilibrium Adsorption of Hg 2+ Ions by Pristine and Functionalized Hly Materials
Batch adsorption experiments were done to establish the adsorption capabilities, the adsorption affinities, and the adsorption mechanisms of p-Hly and of the two functionalized halloysites for Hg 2+ ions. Langmuir and Freundlich equations (all equations are two-parameter models) were used to fit the experimental data at various experimental conditions. The adsorption experiments were carried out over a limited pH range (3-5) in order to minimize the effect of Hg 2+ hydrolysis (see the section above on Speciation Analysis). The background salt and temperature effects were evaluated using 0.1 mol L −1 NaNO 3 or NaCl as ionic media and varying the temperature in the range 283.15-313.15 K. The parameters of the two models are reported in Tables 2  and 3 together with statistical parameters of fits.  The experimental data, together with the fitted curves of the two isotherm models, are reported in Fig. 4 and in Figs S2-S7. The largest R 2 and smallest standard deviations of the fits were obtained using the Langmuir model. For this reason, the sorption abilities and affinities of both the adsorbent materials for Hg 2+ were analyzed using the parameters obtained using the Langmuir equation.
Although the Langmuir model, as well as the other isotherm equation, is an empirical model, the best fit of which does not necessarily mean a specific adsorption mechanism, the adsorption of Hg 2+ ions onto Hly-NH 2 and Hly-SH is a monolayer adsorption due to the presence of the two functional groups grafted onto the external surface of p-Hly (-NH 2 or -SH). Indeed, at the same pH, the adsorption ability of p-Hly is almost zero or, in any case, very small compared to that of functionalized Hly (see Table 2). Moreover, each Hg 2+ ion occupies, hypothetically, a site which is no more available for the other metal ions in aqueous solution.
In general, the functionalization of p-Hly with thiol or amino groups increased its adsorption ability toward Hg 2+ ions in the pH range investigated (e.g. at pH = 4, q m = 4.5, 30.5, and 30 mg g −1 for p-Hly, Hly-SH, and Hly-NH 2 , respectively) ( Table 2, Fig. S2). The experimental conditions for the metal ion solution (pH, ionic medium, and temperature) play an important role in the adsorption process and are summarized as follows: (1) the adsorption capabilities of Hly-NH 2 and Hly-SH at initial pH 3 and 4, and I → 0 mol L −1 are almost the same and in both systems (Hg 2+ ions adsorbent) decrease at pH = 5 with the most pronounced decrease being for Hly-NH 2 material (Table 2, Fig. 5); (2) the addition of 0.1 mol L −1 NaNO 3 to the Hg 2+ solution caused a decrease in adsorption capability for both functionalized halloysites (Table 3, Fig. 6, and Figs S4, S5, and S7);  (3) the addition of 0.1 mol L −1 NaCl to the Hg 2+ solution, done only in the experiments with Hly-SH material, increased its adsorption ability (Table 3, Fig. 6, and Fig. S6); (4) independent of the experimental conditions, the affinity (K L ) of Hly-SH for Hg 2+ ion was greater than that of Hly-NH 2 ; (5) the affinity (K L ) of both Hly-NH 2 and Hly-SH materials for Hg 2+ ion increased with increasing temperature of the metal ion solution (e.g. for Hly-SH, at pH = 3.5 and in 0.1 mol L −1 NaNO 3 , K L = 1.3, 2.5, and 4.1 L mg −1 at T = 283.15, 298.15, and 313.15 K, respectively). Changing temperature caused no significant changes in q m ( Table 3).
The greater affinity and adsorption ability of Hly-SH relative to Hyl-NH 2 was attributed to the greater degree of functionalization (0.4 mmol g −1 and 0.07 mmol g −1 for Hly-SH and HlY-NH 2 , respectively) together with the greater affinity of thiol groups for Hg 2+ ions due to the soft nature of S-donor ligands.
Considering that, at the pH investigated, the amino groups and thiol groups were fully protonated (see the section on Speciation Analysis above), an ionic exchange may, conceivably, have occurred during Hg 2+ adsorption on both functionalized adsorbent materials. Although the H + displacement should be favored by the soft nature of SH groups, at pH ≤ 4, the high H + concentration caused a flattening of the differences in adsorption capability of Hly-SH and Hly-NH 2 , reducing the binding of Hg 2+ ions.
The reduction in adsorption capability of both functionalized halloysites in 0.1 mol L −1 NaNO 3 was justified by assuming, as mentioned above, the competition of sodium cations, which are much more concentrated than Hg 2+ ions. Nevertheless, when the anion of background salt is chloride, at the same ionic strength, an improvement of the adsorption capacity of Hly-SH was found (q m = 30.5, 19.3, 46 mg g −1 at I → 0 and at I = 0.1 mol L −1 in NaNO 3 and NaCl medium, respectively). The same adsorption improvement with increasing NaCl concentration (0 ≤ I (mol L −1 ) ≤ 0.1) was found by Manohar et al. (2002) for Hg 2+ adsorption onto a natural clay impregnated with 2-mercaptobenzimidazole. The formation of Hg 2+ -Cl − species probably changed the speciation picture of the metal ion, increasing the percentages of neutral and negatively charged Hg species (HgCl 2 , HgCl 3 − , HgCl 4 2− ) in solution. In support of this hypothesis, Walcarius and Delacôte (2005) found that the adsorption of positively charged species of mercury (Hg 2+ , HgOH + ) onto thiol functionalized mesoporous silicas reduced their adsorption capability. Those authors attributed this reduction to the progressive charging of the adsorbent material which prevented the adsorption of larger amounts of the toxic metal. As a consequence, the adsorption of neutral species avoided the charging of the adsorbent. A further hypothesis related to this adsorption improvement could be the favorable interaction between protonated -SH groups and negatively charged HgCl 3 − and HgCl 4 − species in the pH range investigated. Considering that, in the pH range investigated, both the -SH and -NH 2 groups of Hly-SH and Hly-NH 2 materials are fully protonated, presumably in proportion to the degree of functionalization, the experimentally observed effect of chloride on the adsorption ability of Hly-SH for Hg 2+ ions should be the same for Hly-NH 2 .
A careful literature search showed that this is the first study dealing with adsorption of Hg 2+ ions onto functionalized halloysite adsorbents. Some rough comparisons can be made with various clay or silica adsorbent materials Hly-NH 2 from aqueous solutions at pH = 3.5/4, in 0.1 mol L −1 NaNO 3 , 0.1 mol L −1 NaCl, and with no ionic medium, at T = 298.15 K functionalized with the same binding groups. Values for q m = 34 and 118 mg g −1 for Hg 2+ ion adsorption onto sepiolite (SEP) and 3-mercaptopropylsylil-sepiolite (MPS-SEP), at pH = 3, were found by Celis et al. (2000). Considering that MPS-SEP has 0.83 mmol g −1 of thiol groups (slightly more than twice that of Hly-SH), the adsorption ability of Hly-SH is comparable (q m = 0.71 and 30.7 mg g −1 at pH = 3, for p-Hly and Hly-SH, respectively) and highlights the important role of thiol groups in Hg 2+ adsorption.
Thermodynamic State Functions of Hg 2+ Ion Adsorption onto Hly-NH 2 and Hly-SH The Gibbs and van't Hoff equations were used to calculate the thermodynamic properties ΔG, ΔH, and ΔS. To this end, the conditional values for K L at pH = 3.5, in NaNO 3 , at I = 0.1 mol L −1 , and in the temperature range 283.15-313.15 K were used (Fig. 7). The ΔG, ΔH, and ΔS values (Table 4) revealed that the adsorption of Hg 2+ ions onto the two functionalized halloysites was a spontaneous process, with negative and very similar ΔG values. Moreover, in both cases, ΔG decreased with increasing temperature from 283.15 to 313.15 K. The adsorption process was endothermic (ΔH = 17 and 28 kJ mol −1 for the Hly-NH 2 and Hly-SH adsorbents, respectively) and caused a small positive entropy variation (ΔS = 0.16 and 0.20 kJ mol −1 K −1 for Hly-NH 2 and Hly-SH adsorbents, respectively). The positive ΔS values, though unremarkable, suggested a structural change in the adsorbent and an increasing randomness at the adsorbent-solution interface (Aksu 2002;Liu and Liu 2008;Tran et al. 2016).
The ΔH values, for both of the adsorbents, at <40 kJ mol −1 , are typical of a physical adsorption mechanism (Gereli et al. 2006;Onal et al. 2007;Liu and Liu. 2008;Tran et al. 2016). In the temperature range investigated, the -ΔG values vary in the ranges 28.7-33.6 and 29.4-35.5 kJ mol −1 for Hly-SH and Hly-NH 2 adsorbents, respectively. ΔG values between -20 and 0 kJ mol −1 are usually ascribed to a physisorption mechanism, whilst free energy values in the range -20--80 kJ mol −1 are considered typical of adsorption based on ion exchange (Gereli et al. 2006;Onal et al. 2007;Tran et al. 2016). The ΔG values, therefore, suggested a physisorption mechanism probably enhanced by an ion-exchange contribution, in agreement with an hypothesis formed on the basis of acid-base properties of functional groups of adsorbent materials and of the speciation studies of Hg 2+ ion in aqueous solution.

CONCLUSIONS
Halloysites functionalized with amino and thiol groups, previously characterized Massaro et al. 2019), were used as adsorbent materials for the removal of Hg 2+ ions from aqueous solutions. The effect of ionic medium, pH, and temperature on the behavior of both adsorbents was evaluated by carrying out kinetic and isotherm experiments under various experimental conditions. The conclusions are summarized as follows: • Thiol and amino groups increase the adsorption capability of halloysite for Hg 2+ ions in the pH range 3-5; • A speciation study of Hg 2+ ions and of adsorbent materials in aqueous solutions at the same experimental conditions as for adsorption experiments was done. Information about the charge and percentage of Hg 2+ ions and of adsorbent species in solution were used to interpret the adsorption mechanisms; • Adsorption equilibria were reached after ~200 min for both Hly-SH and Hyl-NH 2 adsorbents; Ver and the PFO were the best kinetic models for Hg 2+ adsorption onto Hly-SH and Hly-NH 2 adsorbents, respectively; • The Langmuir equation gave the best results in terms of isotherm data fit. The maximum adsorption of Hg 2+ ions by Hly-NH 2 and Hly-SH was obtained in the pH Fig. 7 Plot of lnK L vs. -1/RT for the calculation of thermodynamic state functions ΔH and ΔS for Hg 2+ ion adsorption onto Hly-NH 2 (◻) and Hly-SH (〇) at pH = 3.5, in 0.1 mol L −1 NaNO 3 using the van't Hoff equation range 3-4. The addition of 0.1 mol L −1 NaNO 3 to the Hg 2+ solution reduced the adsorption capability of both adsorbents. The opposite adsorption behavior was found for 0.1 mol L −1 NaCl for Hly-SH; • Thermodynamic state functions (ΔG, ΔH, and ΔS) of adsorption were calculated from equilibrium adsorption data over the temperature range 283.15-313.15 K; • The results obtained suggest a physical adsorption mechanism enhanced by the ion-exchange contribution for both functionalized materials.