3.2 Degradation of 2,4-D photoinduced by H2O2/goethite system in presence of fluoride and bicarbonates at natural concentrations.
Fenton-like, photo-Fenton-like, and photocatalytic processes were carried out with concentrations of goethite (α-FeOOH) often found in natural well waters (0.3 mg L-1) in presence of H2O2 (10 mg L-1), and the effect of natural concentrations of fluoride (1.2 mg L-1) and bicarbonate (86 mg L-1) was evaluated (Figure 5).
2,4-D at concentrations of 30 mg L-1 did not undergo degradation in water containing HCO3-(86 mg L-1) and F- (1.2 mg L-1) by simulated sunlight irradiation (Figure 5: HCO3-/F-/SL), such as we have already reported under UV-B, UV-A, and visible light irradiation, 2,4-D concentration was not strongly reduced in milli-Q water suggesting that photolysis of this molecule did not play an important role.
Although Lin and Gurol claimed that heterogeneous Fenton reaction could take place in goethite/H2O2 systems yielding hydroxyl (OH) and peroxyl (HO2) radicals, the former able to oxidize organic molecules (E°= 2.31 V vs. NHE) and the latter with a lower oxidizing power , dark experiments showed here that in presence of goethite (0.3 mg L-1/0.2 mg L-1 of total iron), H2O2 (10 mg L-1) and anions (F- and HCO3-), 2,4-D concentration was not reduced (Figure 5: G/F-/HCO3-/H2O2/Dark, being G: goethite). In the aforementioned report, authors suggested that, in dark conditions, H2O2 could be adsorbed onto goethite surfaces and undergoes further decomposition yielding OH and HO2 radicals (Eq. 1-3) (heterogeneous Fenton). Low concentrations of goethite used herein would be responsible for the minor role played by heterogeneous Fenton. Moreover, the dark adsorption of 2,4-D on goethite (G/Dark) was found to be negligible.
\begin{equation} H_{2}O_{2}\ \leftrightarrow(H_{2}O_{2})_{s}\ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ (1)\nonumber \\ \end{equation}\begin{equation} \equiv\text{Fe}^{3+}-OH+(H_{2}O_{2})_{s}\ \leftrightarrow\ \ \ \ \ \ \ \text{Fe}^{2+}+H_{2}O+HO_{2}^{\bullet}\text{\ \ }\ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ (2)\nonumber \\ \end{equation}\begin{equation} \equiv\text{Fe}^{2+}+\ H_{2}O_{2}\ \rightarrow\ \equiv\text{Fe}^{3+}-OH+^{\bullet}\text{OH}\ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ (3)\nonumber \\ \end{equation}
Photocatalytic reactions under sunlight irradiation (in absence of H2O2) of goethite in the presence or absence of anions led to a 2,4-D degradation of 10% after 240 min (Figure 5: G/SL; G/F-/HCO3-/SL). As it was described above, α-FeOOH does not show an important photocatalytic behavior since its conduction band potential is very positive and unable to reduce molecular oxygen leading to the high e-/h+ pair recombination despite some authors have argued that goethite seems to be a promising photocatalyst removing several organic pollutants such as dyes, chlorophenols, and polycyclic aromatic hydrocarbons . However, in these aforementioned studies, goethite concentrations ranged from 1 to 500 g L-1 (herein it was used a concentration of 0.0003 g L-1) and the high pollutant removal rates observed may probably be directly related to high photocatalyst concentrations. In addition, the photocatalytic reaction did not change in the presence of bicarbonate and fluoride.
Otherwise, when 10 mg L-1 of H2O2 was added into water containing goethite (photo-Fenton-like/photocatalytic reactions) and bicarbonate at initial pH 6.9 under simulated sunlight irradiation, 2,4-D concentration was strongly reduced (75%) after 240 min of simulated sunlight irradiation (Figure 5: G/HCO3-/H2O2/SL). Several authors have claimed the positive effect of H2O2 addition into α-FeOOH photocatalytic systems to degrade several pollutants . Nevertheless, these studies were performed by using α-FeOOH and H2O2 concentrations of 1 g L-1 and 50-170 mg L-1 respectively which are much higher than natural amounts of goethite (0.0003 g L-1) and H2O2 (10 mg L-1) used herein.
On the other hand, under sunlight irradiation, surface\(\equiv\text{Fe}^{2+}\)species could be photoinduced (Eq.4) . These species could react with H2O2 leading to the generation of surficial \(\equiv\text{Fe}^{3+}\)and extraOH radicals (heterogeneous photo-Fenton) (Eq. 4).
\begin{equation} \equiv\text{Fe}^{3+}-OH+hv\ \rightarrow>\equiv\text{Fe}^{2+}+\ {\text{OH\ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ }\left(4\right)}\ \nonumber \\ \end{equation}
Furthermore, surficial \(\equiv\text{Fe}^{3+}-OOH\) species generated by H2O2 adsorption on goethite surface could (Eq. 5), by UV-vis light irradiation, to photo-induce ferryl species (\(\equiv\text{Fe}^{4+}=O)\), this latter very unstable in aqueous solutions, also leading to the formation ofOH radicals (Eq. 7) .
\begin{equation} \equiv\text{Fe}^{3+}-OH+\ H_{2}O_{2}\rightarrow\ \equiv\text{Fe}^{3+}-OOH+\ H_{2}O\ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ (5)\nonumber \\ \end{equation}\begin{equation} \equiv\text{Fe}^{3+}-OOH+hv\ \rightarrow\equiv\text{Fe}^{4+}=O+^{\bullet}\text{OH}\ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ (6)\nonumber \\ \end{equation}\begin{equation} {\ \equiv Fe}^{4+}=O+\ H_{2}O\ \rightarrow\ \equiv\text{Fe}^{3+}-OH+\ ^{\bullet}\text{OH}\ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ (7)\ \nonumber \\ \end{equation}
Bicarbonates (83.6 mg L-1) which were also present at typical concentrations often found in natural waters regulating the pH at 6.9, could react with OH radicals generating carbonate radicals (CO3-•) (Eq. 8) able to remove 2,4-D as well . CO3-•radicals could participate in electron transfer and H-abstraction reactions leading to the oxidation of organic molecules . All these routes mentioned above could be responsible for the significant removal of 2,4-D in presence of bicarbonate.
\begin{equation} \text{HCO}_{3}^{-}+\ ^{\bullet}\text{OH}\ \rightarrow\ \text{CO}_{3}^{-\bullet}+\ H_{2}O\ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ (8)\nonumber \\ \end{equation}
Surprisingly, experiments, where only hydrogen peroxide, fluoride, and bicarbonate were present (in absence of goethite) at concentrations of 10, 83.6, and 1.2 mg L-1 respectively under simulated sunlight irradiation at initial pH 6.9, exhibited a strong 2,4-D removal of around 85% after 240 min (Figure 5: H2O2/F-/HCO3-/SL). As it has been suggested, sunlight or UV-B light irradiation can photo-induce H2O2 photolysis (Eq. 9) . Pyrex-glass reactors used can allow the transmission to a lesser extent of UV-B light coming from the solar simulator leading to the H2O2 photolysis.
\begin{equation} H_{2}O_{2}+hv\ \left(280-315\ nm\right)\ \rightarrow 2^{\bullet}\text{OH}\ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ (9)\nonumber \\ \end{equation}
The experiments of H2O2 photolysis without anions (H2O2/SL) revealed that 2,4-D degradation underwent a slight reduction. Perhaps, bicarbonate reaction with OH radicals leading to the generation of CO3-• could be behind the observed enhancement (Eq. 8).
The photocatalytic experiment in absence of fluoride (G/HCO3-/H2O2/SL) showed evidence about goethite could affect the OH radical production by H2O2 photolysis. Two phenomena could be responsible for this detrimental effect: (i) H2O2 oxidation (+1.06 V vs NHE) by photoinduced goethite valence band holes (VB oxidation potential: +2.3 V vs NHE ) could also take place producing a low oxidant peroxyl radical (HO2) (Eq. 10; Figure 7) unable to oxidize 2,4-D. Peroxyl radical is in equilibrium with superoxide radical (O2-•) (Eq. 11) and its pKa is around 4.8, so at pH higher than pKa (pH during experiments increased from 6.9 to 7.5) superoxide radical will be the main specie in solution which undergoes a fast disproportion leading to the generation of H2O2 (Eq. 12-13) [43].
\begin{equation} {H_{2}O_{2}+\ h_{\text{VB}}^{+}\ \rightarrow\ \text{HO}_{2}^{\bullet}+H^{+}\ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ (10)\backslash n}{\text{HO}_{2}^{\bullet}\ \leftrightarrow\ H^{+}+\ O_{2}^{-\bullet}\text{\ \ \ \ \ \ \ \ \ \ \ \ p}K_{a}=4.8\ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ (11)}\nonumber \\ \end{equation}\begin{equation} \text{HO}_{2}^{\bullet}+\text{HO}_{2}^{\bullet}\ \rightarrow\ H_{2}O_{2}+\ O_{2}\ k=8.6x10^{5}M^{-1}s^{-1}\ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ (12)\ \ \ \ \ \ \ \ \ \ \nonumber \\ \end{equation}\begin{equation} O_{2}^{-\bullet}+\ O_{2}^{-\bullet}+\ {2H}^{+}\rightarrow\ \ H_{2}O_{2}+\ O_{2}\ k=\ll 100\ M^{-1}s^{-1\ }\ \ \ \ \ \ \ \ \ (13)\ \ \ \nonumber \\ \end{equation}
On the other hand, (ii) goethite exhibits an important UV light absorption (Figure 3b) generating possibly a screen effect and competing with H2O2 molecules by UV-B photons.
Another pathway for ROS generation could be H2O2 direct reduction by goethite CB photoinduced electrons. Since hydrogen peroxide reduction potential is +0.32 V (vs NHE) and photoinduced conduction band electrons in goethite may have a redox potential of +0.24 V (vs NHE) [44], it is expected that H2O2 reduction by goethite CB photoinduced electrons takes place leading to the generation of hydroxyl radicals (eq. 14).
\begin{equation} H_{2}O_{2}{\ +\ e}_{\text{CB}}^{-}\rightarrow\ ^{\bullet}{OH+\ \text{OH}^{-}\text{\ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ }\left(14\right)}\nonumber \\ \end{equation}
As was expected, when natural amounts of fluoride (1.2 mg L-1) were added to water containing already bicarbonates (83.6 mg L-1) in presence of goethite and H2O2 (10 mg L-1), the 2,4-D degradation increased markedly reaching 95% in 240 min of simulated sunlight irradiation (Figure 5: G/F-/HCO3-/H2O2/SL). XPS measurements of goethite separated from the aqueous solution after the reaction confirmed the presence of fluoride on goethite (1.2 %At) (Figure 6a) and a high-resolution F 1s spectrum (Figure 6b) revealed the presence of two signals at 684.7 eV and 689.5 eV often linked to the presence of adsorbed fluoride in iron (hydr)oxides and Fe-F bonds respectively .
Hiemstra and Van Reimsdijk concluded that fluoride adsorption onto goethite is achieved on several surface sites being benefited at acidic pH values (goethite isoelectric point 7.9). Moreover, the authors suggested by theoretical calculations that the formation of surface Fe-F bonds would be feasible. On the other hand, Ding et al. corroborated that the highest fluoride adsorption was obtained at acid pH values and proposed that fluoride adsorption may be achieved by exchange with surficial -OH groups. In other ways.
As it was above mentioned, the beneficial effect of fluoride presence over photocatalytic degradation using goethite and lepidocrocite was first claimed by Du et al. . The use of 8.5-85 mg L-1fluoride increased the photocatalytic degradation of orange II dye at pH 6.5 by irradiation of these iron (hydr)oxides (0.5 g L-1). Authors suggested that fluoride could modify the surface of iron (hydr)oxides increasing the production ofOH radicals, however, the mechanism was not proposed. Meanwhile, on TiO2 a mechanism was suggested. In the literature there are several studies suggesting different mechanisms about how fluoride anions could enhance the photocatalytic activity of metal oxide semiconductors as TiO2. For instance, Minella et al., suggested that the flat band potential of semiconductors may be fine-tuned by surficial anion adsorption. This flat band modification will depend on the nature and density of ions. Otherwise, Minero et al. , indicated that fluoridation of TiO2could enhance the photocatalytic activity of the metallic oxide towards the phenol degradation due to the formation of Ti-F surface sites. Authors suggested that the exchange of surficial -OH groups by F atoms on TiO2 may decrease the formation of deeply trapped holes, favoring less deep surface trapping sites and making faster the electron transfer of photoinduced holes to phenol molecules.
In addition, a systematic study reported by Deskins and coworkers where DFT calculations were carried out, demonstrated that rutile TiO2 surfaces could be strongly modified by adsorbates (especially those molecules highly electronegative) tailoring the surface chemistry. Later, Montoya and Salvador measured the flat band potential of fluorine-modified TiO2 finding that this underwent a negative shifting of about 80 mV compatible with upward band bending in semiconductors by an excess of negative charge in the surface. For their part, Xu et al. proposed a new theory where they suggested that fluoride ions present in the Helmholtz layer can induce the desorption of surface-bound OH radicals photoinduced in TiO2 surfaces through a fluorine hydrogen bond (Figure 6). This latter could support our results where the presence of two different species of fluoride interacting with the goethite surface were found by XPS.
Other authors have also suggested that the presence of surficial\(\equiv Ti-F\) species may play an important role since photoinduced VB holes may be unable to oxidize the F-(E(F/F-)=3.6 V vs NHE) and would react directly with water molecules yielding freeOH radicals .
\begin{equation} \equiv Ti-F+\ h_{\text{VB}}^{+}\rightarrow\ \equiv Ti-F+\ ^{\bullet}\text{OH}+\ H^{+}\ \ \ \ \ \ \ \ \ \ (15)\nonumber \\ \end{equation}
Thus, it is possible suggesting that α-FeOOH surfaces in presence of fluoride (at natural concentrations) can be superficially modified leading to either/both upward band bending or/and enhanced generation of free hydroxyl radicals upon sunlight irradiation (Figure 7). Upward band bending could enhance the H2O2 reduction by photo-induced conduction band electrons in goethite (Eq. 14) yielding more efficiently OH radicals, therefore, enhanced photocatalytic degradation of 2,4-D, such as was observed.
The effect of homogeneous photo-Fenton reactions was also assessed since often iron (hydr)oxides can undergo photoinduced iron dissolution in water . Factors such as acidic pH and the presence of aliphatic acids as oxalate (by forming soluble ferric-oxalate complexes at circumneutral pH, for this reason, herein photocatalytic experiments were not carried out in absence of bicarbonates, since under these conditions, the pH of the solution dropped reaching values about of 4.0 promoting goethite dissolution) can also enhance iron dissolution. The presence of dissolved iron either complexed or not can induce homogeneous photo-Fenton reactions in presence of hydrogen peroxide . In our experimental conditions, the final pH of the reaction was always around 7.5 confirming that 2,4-D removal was achieved at neutral/basic pH. The presence of dissolved iron was evaluated by atomic absorption spectroscopy measurements of the supernatant obtained by further 0.22 µm filtration after the photochemical and Fenton-like reactions. In dark experiments, the total iron concentration was 2.81 µg L-1 while the simulated sunlight-irradiated experiment in absence of hydrogen peroxide showed an iron concentration of 2.9 µg L-1. The simultaneous presence of goethite, anions (F- and HCO3-), H2O2, and simulated sunlight exhibited an iron concentration of 11 µg L-1. These iron amounts are too low to induce Fenton or photo-Fenton reactions at neutral pH able to generate appreciable concentrations of OH radicals responsible for herbicide removal; pointing out that the degradation was mainly due to heterogeneous photo-Fenton processes/sunlight H2O2 photolysis.