Materials

Synthesis of LDHs. Carbonate-containing Mg-Al LDHs with a Mg/Al atomic ratio close to 3 were prepared by a sol-gel method. A dissolution of aluminium tri-sec-butoxide (ATB, Aldrich 99.9 %) in ethanol was refluxed and stirred for 1 h. Afterward the temperature was decreased to 0 °C, 3 M HNO₃ was dropped, and the mixture was stirred 1 h. Following this, magnesium methoxide (Aldrich, 99%) dissolved in butanol, and water were slowly dropped into the solution until a gel was formed, which was dried at 70 °C. In the synthesis of fluorinated LDHs, a part of ATB was replaced by sodium hexa-fluoroaluminate, Na₃AlF₆ (Aldrich, 99.99%). The ratio of Mg/Al was maintained at 3; the fluorine content was varied according to Table 1. The chemical composition reported in Table 1 is the result obtained from thermal analysis and chemical analysis conducted by inductively coupled plasma-mass spectrometry (ICPMS), where a Thermo Scientific™ ELEMENT 2™ system was used.

Characterization. The LDH samples were characterised by X-ray diffraction (XRD), thermal analysis (TGA), infrared (FTIR) spectroscopy and dye adsorption followed by optical spectroscopy.

The XRD patterns were acquired using a diffractometer (D8 Advance-Bruker) coupled to a copper anode X-ray tube.

Mid-infrared (FTIR) spectra were acquired at room temperature using a Perkin Elmer Series spectrophotometer (Model 6X) operated in the ATR-FTIR mode. The spectra were recorded over the 400-4000 cm^-1 spectral window by averaging 32 scans at a resolution of 4 cm^-1.

The thermogravimetric analyses were carried out using a thermobalance Q500HR (TA Instruments).

The evaluation of dipolarity/polarisability (parameter π*) of samples was determined by monitoring solvent-dependent shifts of the UV/vis absorption band of 4-tert-butyl-2-(dicyano-methylene)-5-[4-(diethylamino)benzylidene]-Δ3- thiazoline as the probe dye by means of the multi-parameter method of Kamlet-Taft [²³-²⁵] where π* is estimated from the UV/vis absorption maxima of adsorbed dye according to: π* = 9.475−0.54ν_max [10⁻³ cm⁻¹]. A spectrometer Perkin-Elmer Lambda 40 was used to acquire the UV-vis powder spectra in reflection mode. Dye containing samples spectra were recorded with dye-free powders as a reference.

THMs uptake by layered double hydroxides. All the sorption experiments were performed in batch at room temperature (22°C), using fresh or calcined LDHs. The ratio mass of adsorbent to volume of solution was fixed to 12.5 mg/ml. For each experimental run, 10 ml aqueous solution of the THM was taken in a 50ml vessel containing 0.125 g of the adsorbent. Suspensions were agitated at 150 rpm. At the end of the predetermined time, t, the suspensions were centrifuged, and the supernatant analysed for THM concentration by ¹H NMR. ¹H NMR spectra were acquired in a Bruker Avance 400 NMR spectrometer, operating at 400 MHz. The samples were loaded in a 5 mm NMR tube using CDCl₃ with TMS reference as solvent. The signal at 7.2 ppm, corresponding to chloroform was integrated and the concentration calculated by comparing to integrated signal of reference. The same method was applied to monitoring CHBr₃, integrating the signal at 6.8 ppm.

Adsorbents

X-ray diffraction patterns of the fresh samples are displayed in Fig. 1. In three samples the hydrotalcite phase (JCPDS card 22-0700) was indexed and no crystalline fluorine compounds were identified. The degree of crystallization is low as frequently is observed in samples prepared by sol-gel. Because of the broadness of XRD peaks is hard to confirm any influence of the presence of fluorine on the position of (003) peak. Earlier [²⁶], in samples of LDH prepared by sol gel was demonstrated that samples exhibited broad XRD peaks the total amount of Al is incorporated to the LDH lattice and consequently brucite is not formed.

Fig. 1 X-ray diffraction patterns of LDH as synthesized: (a) HT, (b) HT-10F and (c) HT-25F. Peak labels indicate Miller index of JPCDS card 22-0700. 

Fig. 2 shows the FTIR spectra of layered samples where the absorption band due to the O−H stretching is observed from 3700 to 3000 cm⁻¹. The intensity of this band is modified with the presence or the absence of fluorine. Actually, infrared intensities of the broad OH stretching bands were integrated by using the EZ OMNIC 32 software. The integrated intensities were 1310, 1223, and 1011 for HT, HT-10F, and HT-25F, respectively. These values support the replacement of OH^- by F^-, as previously reported in the synthesis of fluorinated LDH by coprecipitation. The band attributed to the bending deformation of molecular water is observed at 1650 cm⁻¹. The band assigned to the asymmetric stretching of C−O bonds in carbonate ions [²⁷-²⁸] is observed between 1340 and 1450 cm⁻¹. A fact that should be emphasized is the presence of a band at 1554 cm^-1 which is assigned to monodentate carbonate species, this band is stronger in spectra HT and HT-10F than in spectrum of sample HT-25F. This monodentate carbonate species have been related with the presence of strong basic sites [²⁹]. Thus, it seems that fluorination of brucite-like layers have eroded strong basic OH^- sites, since some of this OH^- groups were changed by F^-.

Fig. 2 Infrared spectra of LDH as synthesized: (a) HT, (b) HT-10F and (c) HT-25F. 

Fig. 3 displays representative UV/vis absorption spectra of dye adsorbed on the thermal treated LDH samples. The position of absorption band is dependent of the composition of sample, the higher the fluorine content the higher the wave length. From these spectra, the π* values for solids HT, HT-10F, and HT-25F were 0.99, 1.18 and 1.27, respectively. Clearly, the presence of fluorine increases the dipolarity/polarisability at surface of LDH which is explained because the different electronegativity of F and O atoms leads to formation of dipoles differing in magnitude and orientation. The replacement OH^- by F^- is aleatory trough the bulk and surface of LDH. However, the presence of fluorine content affects the stacking of layers leading to materials differing in morphology as shown in SEM images in Fig. 4. The particles of sample without fluorine have an irregular shape. Small particles agglomerates lying on top of one another can be observed to form particles as big as 80 μm. In contrast, in samples containing fluorine the layers are aggregated to form particles smaller than 40 μm.

Fig. 3 UV-vis spectra of the solvatochromic dye 4-tert-butyl-2-(dicyano-methylene)-5-[4-(diethylamino)benzylidene]-Δ3-thiazoline adsorbed onto (a) HT, (b)HT-10F and (c) HT-25F. 

Fig. 4 SEM image of (a) HT, (b) HT-10F and (c) HT-25F. The bar scale in image (a) corresponds to 20 μm and it is valid for all three images. 

In summary, the fluorination of brucite-like layers leads to materials with different dipoles at surfaces prone to stabilize dipolar and polarizable molecules. So, the following section is intended to explore the potential retention of trihalomethanes in fluorinated LDH trough dipolar interactions sorbate-adsorbent.

Thrihalomethanes adsorption

Adsorption equilibrium time. The time-dependent curves of THM solution concentration in contact with different adsorbents are shown in Fig. 5. The concentration of CHCl₃ decreases significantly in 10 first minutes of contact with adsorbent. The major decreasing in solution concentration during first 10 minutes was observed when the adsorbents used were the samples containing fluorine. Further, it is clear that fresh samples adsorb less THM than the thermal treated ones. After ten minutes, the adsorption on LDHs continues slowly and after 20 minutes it can be assumed that the equilibrium CHCl_3(aq) + LDH_(sol) ↔ (LDH.xCHCl₃)_(sol) + (1-x)CHCl_3(aq) was reached.

Fig. 5 Concentration progress of CHCl₃ (left) and CHBr₃ (right) after contact with different adsorbents (a) HTC, (b) HT-10FC and (c) HT-25FC. Initial concentration of CHCl₃ and CHBr₃ was 5X10^-3 and 4X10^-4 M, respectively. Adding of C to code sample means that adsorbent was thermal treated at 350 °C for 8 h. The ratio mass of adsorbent to volume of solution was fixed to 12.5 mg/ml. 

The adsorption experiments made with CHBr₃ progressed slower than that with CHCl₃. Again, the curves show a first period of 10 minutes were the solution concentration diminishes significantly but, differently to CHCl₃, the concentration continues decreasing for periods as long as 90 minutes, thus this time was taken as the equilibrium time. Even when the initial concentrations of CHCl₃ and CHBr₃ were not the same, results in the next sections will confirm that nature of the THM is very influent in the adsorption progression. The different concentrations were taken because of the different solubility of THM and the goal in this study to explore removal of THM at low concentration.

The effect of the composition and structure of adsorbent. In Fig. 6 is plotted the percent of CHCl₃ removal trough adsorption onto different adsorbents. Clearly, the solids thermally treated previous to adsorption remove the major fraction of chloroform which is explained because the presence of numerous polar sites present at surface of thermally treated samples [²²]. This trend is observed also when the adsorbat is CHBr₃. Thus, the high specific surface area and major number of polar sites in adsorbate enhance the adsorption of CHCl₃ and CHBr₃.

Fig. 6 CHCl₃ (left) and CHBr₃ (right) removal efficiency through adsorption on different adsorbents. The equilibrium time was taken as 30 and 90 minutes for CHCl₃ and CHBr₃, respectively.  

Independently of the adsorbent, fresh or thermal treated, the presence of fluorine in adsorbents clearly increases the capacity of adsorbed THM. For example, the sample HT25FC removes 97% of CHCl₃ in solution but the free-fluorine adsorbent only eliminates 33% of CHCl₃ initially present in solution.

Another highlighted observation is that percentage of CHCl₃ removed by adsorption on LDH is higher than that observed for CHBr₃. The dipole moment of CHCl₃ (1.122D) is slightly higher than that of CHBr₃ (0.99 D). Further, CHBr₃ is bigger than CHCl₃, thus, this differences limit the dipole interactions with polar sites at surface of adsorbents.

The effect of the initial THM concentration solution. The adsorption of CHCl₃ onto LDHs was measured at different CHCl₃ concentrations in the range from 5x10^-3 to 5x10^-2 M. Results are presented in Fig. 7. It is clear that the initial concentration is a significant driving force to prevail over all mass transfer resistance of CHCl₃ between the aqueous and solid phases. Thus, a higher initial concentration of CHCl₃ increased adsorption of CHCl₃. This result was consistent for shortest periods (first 10 minutes) and for all adsorbents. Some remarks have to be enumerated: 1) the adsorbent without fluorine is the one with the lowest adsorption capacity. In this material series an initial CHCl₃ concentration of 5x10^-3 M leads to a removal of CHCl₃ through adsorption as low as 28%. With increase of CHCl₃ concentration the adsorptivity increases until removal 58% of THM when the initial concentration is 0.05 M. 2) the rate of adsorption and the adsorption capacity of fluorinated adsorbents is significantly higher than that of the free-fluorine adsorbents. The adsorbent HT10FC removes after 20 minutes 78, 82, 89 and 95 % of the CHCl₃ when the initial CHCl₃ concentration is 5x10^-3, 2x10^-2, 3x10^-2 and 5x10^-3 M, respectively. On the other hand, the adsorbent with the highest loading of fluorine is the best adsorbent, taking in the first 10 minutes more than 94% of CHCl₃ present initially in solution, independently of the initial CHCl₃ concentration. 3) The high capacity of adsorption showed by fluorinated LDHs suggest that number of adsorption sites is modified with the fluorination. Further, the saturation of adsorption sites does not occur as the concentration of CHCl₃ is low in order to reach CHCl₃ concentrations measured in drinking and swimming pool water.

Fig. 7 CHCl₃ adsorption onto different adsorbents (squares) HTC, (circles) HT-10FC and (triangles) HT-25FC, starting with different initial CHCl₃ concentration, 5x10^-3 M (solid lines), 2x10^-2 (dashed lines), 3x10^-2 (dotted lines), 5x10^-2 (dotted-dashed lines). 

In Fig. 8 are contained the results of CHBr₃ adsorption on un-fluorinated and fluorinated LDHs as a function of initial CHBr₃ concentration. In line with the behavior observed with CHCl₃, the percent of CHBr₃ retained on weight unit of adsorbent increase with the increasing of initial CHBr₃ concentration from aqueous solution. Starting from CHBr₃ 4x10^-4 M and 5x10^-3 M the best adsorbent was HT25FC adsorbing close to 70% and 90% of CHBr₃, respectively. For the same initial CHBr₃ concentration, the efficiency of CHBr₃ removal follows the order: HT25FC > HT10FC > HTC. However, it is clear that CHBr₃ adsorption proceeds less efficient than that observed for CHCl₃. As shown in previous section, the CHBr₃ adsorption proceeds slowly than CHCl₃ adsorption. Thus, it seems that this trend is determined by the fact that at higher concentrations, the most superficial groups are already occupied and in consequence, the diffusion of CHBr₃ to the un-reacted functional groups is inhibited. Note that, saturation of adsorption sites is not an alternative explanation as with CHCl₃ adsorption enhanced with concentrations as high as 0.05M. Thus, the effect observed in CHBr₃ should be attributed to the size of molecule inhibiting motion during adsorption.

Fig. 8 CHBr₃ adsorption onto different adsorbents (squares) HTC, (circles) HT-10FC and (triangles) HT-25FC, starting with different initial CHCl₃ concentration, 4x10^-4 M (solid lines), 5x10^-3 (dashed lines). 

Adsorption isotherm. The isotherms plotted in Fig. 9 point out the effect of the presence of fluorine in adsorbents. Linear profiles, q_e=kCe, are obtained when adsorbents do not contain fluorine, which is often observed in cases where adsorbat concentration is low. Interesting, the isotherm of both CHCl₃ and CHBr₃ match to that of Freundlich’s model, q_e=K(Ce)^1/n, when the adsorbent contains fluorine. Further, in both Freundlich isotherms 1/n>1, which means that THM adsorbs forming a mono layer onto adsorbent. Currently, the observed isotherms are obtained for polar compounds. Both CHCl₃ and CHBr₃ have a dipolar moment and, at low concentrations, such compounds are in competition with water for adsorption sites of adsorbent. Of course the adsorption sites should be also polar groups. Clearly, the adsorption of THM is more favorable in fluorinated adsorbents and this feature is explained because of the major polarity induced by the presence of fluorine.

Fig. 9 Adsorption isotherms of CHCl₃ (left) and CHBr₃ (right) onto (a) HTC, and (b) HT-25FC. q_e represents the amount of THM per unit mass of adsorbent and Ce the equilibrium concentration of the remaining THM in solution. The equilibrium time was taken as 30 and 90 minutes for CHCl₃ and CHBr₃, respectively.