3.1 TiO2 and Fe-TiO2 nanoparticles characterization
In Fig. 3 SEM images of the titania nanoparticles are reported. The synthesized nanoparticles appear spherical with an average size of 62.76 nm ± 16.74 nm (Fig. 3a) and with a homogeneous distribution through the matrix. The addition of a certain amount of Fe (0.5 wt %) not produced a morphological modification[14] but has a suppressive effect on the growth of TiO2 crystals (39.88 nm ± 5.51 nm), since the additives hindered the contact between TiO2 particles and inhibited crystal growth during heat treatment[4] and, how revealed in figure (Fig. 3b), conferred an increase on the conductivity of the material as demonstrated by the increase in brightness in the image.
With the increase of the Fe content, a loss of the spherical shape was detected, and the nanoparticles appeared more agglomerated (Fig. 3c-d).
3.2 UV-vis diffuse reflectance spectra
The results of UV-Vis-DRS analysis on Fe-TiO2 and Fe-P25 are shown in Fig. 4 and Fig. 5. The pure TiO2 has a bandgap of 3.05 eV which is consistent with the bandgap of pure anatase (3.2 eV). The diffuse reflectance spectra of all Fe-TiO2 samples exhibited a redshift (to the left). The addition of Fe increased absorption in the visible-light range leading to a decrease in the energy bandgap (Fig. 4, Table 1) thanks to Ti4+ substitution by Fe3+ in the TiO2 lattice that forms bands located near the bottom of the conduction band [16].
As the iron content increases the left side of the bell tends to rise; another distinct phase (an iron oxide) is formed. In the case of the 2%Fe-TiO2 sample, the formation of a third phase can be seen due to the presence of a third peak (slightly lower than that of titania), probably another iron oxide.
These results revealed that the iron ions are indeed incorporated into the lattice of TiO2 because the ionic radius of Ti4+ (0.68A˚) and Fe3+ (0.64A˚) are almost the same and Fe3+ ions can enter the crystal structure of titania [17, 18, 14, 15, 16].
Bandgap energy was estimated from a plot of (ahv)2 vs photon energy (hv). The intercept of the tangent to the plot will give a good approximation of the bandgap energy for TiO2 and the results are reported in Table 1.
Table 1
The bandgap of samples P25, TiO2 and with different iron content: The bandgap energies decrease with increasing Fe-doping concentration.
Sample | Bandgap (eV) |
P25 | 3.20 |
0.5%Fe-P25 | 3.05 |
1%Fe-P25 | 3.00 |
2%Fe-P25 | 2.82 |
TiO2 | 3.05 |
0.5%Fe-TiO2 | 2.81 |
1%Fe-TiO2 | 2.52 |
2%Fe-TiO2 | 2.41 |
For the P25 sample, the doping did not have the same effect, the modification is not as effective as for TiO2. It is possible to observe a shift of the peak towards lower values with the addition of iron, but without a significant modification in the values that remained almost constant for all three samples (0.5%Fe-P25, 1%Fe-P25 and 2%Fe-P25). With the increase of the iron content, the bell widens and the left branch tends to rise, so the iron oxide phase (probably Fe2O3 [14]) is increasing, and migration into the bulk of titanium dioxide is not occurring (Fig. 4, Table 1).
3.3 BET surface areas
The surface area, pore volume and pore size values from the BET analysis are summarized in Table 2 where it is possible to observe that the decrease of the surface area as the iron content increases for the P25 samples is attributable to pore occlusion while the synthesized TiO2 has an opposite behavior; in fact, as the iron content increases, the surface area increases probably because of the iron oxides nanoparticles can alleviate thermal sintering of the TiO2 nanoparticles [14] as supported by SEM investigations showing than Fe-TiO2 particles are smaller than pure TiO2.
Table 2
Surface area, pore volume and pore size values from the BET analysis.
Sample | BET Surface Area (m2/g) | Pore Volume (cm3/g) | Pore Size (Å) |
P25 | 52.69 | 0.12 | 70.83 |
0.5%Fe-P25 | 44.19 | 0.29 | 314.40 |
1%Fe-P25 | 40.75 | 0.32 | 293.88 |
2%Fe-P25 | 36.07 | 0.24 | 254.66 |
TiO2 | 4.18 | 0.01 | 61.50 |
0.5%Fe- TiO2 | 8.01 | 0.03 | 64.24 |
1%Fe- TiO2 | 9.16 | 0.03 | 64.34 |
2%Fe- TiO2 | 10.86 | 0.03 | 50.62 |
Figure 5 shows the N2 adsorption-desorption isotherms for all samples. It can be observed that pure P25 is a mesoporous material while the addiction of iron occludes the pores.
The adsorption isotherm for the sample P25 is a type IV isotherm (according to the IUPAC classification of adsorption isotherms) with a hysteresis cycle; instead, the adsorption isotherms for the samples with iron are type III isotherms, typical of poor interactions between adsorbate and adsorbent material. This result is consistent with that of the UV-vis-DRS analysis: iron does not migrate into the crystalline bulk of P25 but forms another phase that settles on the surface of the nanoparticle occluding its pores [19].
For the TiO2 sample, iron into the titania lattice modifies the material. The adsorption isotherm for the pristine TiO2 is a type III curve, while adsorption isotherms for the samples with iron are type IV curves, characteristic of mesoporous materials [21] (Fig. 5).
3.4 Photocatalytic degradation of MB
All the process parameters were tested one by one to find the optimum condition for our system.
3.4.1 Iron content
Preliminary experimental tests under visible radiation revealed that the adoption of a Fe content in the range (0.5–1.5 wt%) produced and increase in photocatalytic efficiency (Fig. 6) as suggested by the enhancement on MB removal from 35% of pure TiO2 up to 57% (Fe amount of about 0.5 wt%, 1 wt% and 1.5 wt%). This is because the dopant metal replaces TiO4+ in the crystalline bulk of TiO2 bringing improvements in the photocatalytic activity of titania. Fe3+ metal cations can act as photogenerated positive hole-electron pairs [22] having the characteristic of both charge donors and acceptors (Fe2+/3+). All these cations were used in low concentrations (typically between 0.5%- 3% in moles) because if the doped metal ions concentration increases too much, the photocatalytic activity will decrease since the dopant could act as a recombination site [3] and in addition, the formation of the second phase could be observed (for Fe2O3 it occurred at a charge above 2% [14]).
Employing the synthesis method proposed in paragraph 2.2 a Fe loading of 1 wt% allows a suitable dispersion of iron species.
A further increase in Fe content (1.5%) not produced an improvement in MB removal and this can be attributable to the reduction of Ti content in the resulted catalyst and to the effect of Fe3+ ions which can act as charge recombination centers, that hinder the formation of radical species and subsequent oxidation of pollutant molecules [14, 16, 17, 19].
Removal tests also show that TiO2 achieves higher removal than P25. This result can be attributed to the presence of more structural defects in the TiO2 synthesized than P25 (hydrogen inclusion, anion vacancies, reduced Ti species, atom displacements from pure phase positions, etc.) as already observed in another study [11].
As a result of this preliminar tests, subsequent tests were performed with 1%Fe-TiO2.
3.4.2 pH
To optimize the operative conditions during the photocatalytic tests, the effect of MB solution pH was analyzed in the range from 2.2 to 10 and the results are shown in Fig. 7.
The optimal pH condition was found to be near to neutral one (pH = 7): lower and higher values that 7 produced a detrimental effect on MB removal. This is because, at acidic pH conditions catalyst surface could be surrounded by positive charges of H+ and, therefore, repels cationic MB molecules, inhibiting them from reaching active sites of TiO2. Under alkaline conditions (pH > 7) the cationic MB molecules might be covered by negative charges, which repel the negatively charged catalyst surface [20].
3.4.3 Pollutant concentration
As regards the MB initial concentration, as expected, an increase in MB concentration produced a slight decrease in MB removal from 5 ppm to 15 ppm (Fig. 8a) thus suggesting a limiting effect on radicals availability and a delay on MB degradation. At the higher concentration (30 ppm) the degradation was significantly penalized and this because more free radicals and oxidant species for degradation are required and the dark color of the solution affects light penetration [20, 21]. In addition, a large amount of dye is adsorbed on TiO2 particles which are prevented from dye molecule reaction with free radical and electron-holes [22]. At 50 ppm of MB concentration, no degradation was measured because of the prevention of photocatalytic reaction due to no light penetration.
3.4.4 Catalyst dosage
The MB removal increased with the increase of catalyst dosage in the first hour of tests. After that, a plateau on MB removal (45% after 30 min and 47% after 180 min) with a catalyst loading of about 2 g/L was recorded while, in the other tests, the MB removal was ensured in the entire investigated time (Fig. 9a) and the optimum was reached by using 1 g/L as catalyst loading. The increase of the catalyst dosage promotes the increase in the availability of the active sites [15, 21]. The increase in MB removal at 2 g/L in the first 60 min can be attributable to the only adsorption mechanism and, even in the case of light activation, the recorded removal was the result of the achievement of equilibrium conditions [1]. The presence of a high dosage of a catalyst than 1 g/L turned the solution to be turbid, and this inhibits light penetration [1, 20–22]. To support this evidence, a 1 h of test in dark was conducted and the same MB removal was calculated (about 43%).
3.4.5 Irradiance
The effect of Irradiance is reported in Fig. 10 and, as expected, the increase in the light irradiance promoted the MB removal (Fig. 10a) and at the highest irradiance value adopted the maximum MB removal was calculated (80%).
3.5 Occurrence of radical mechanism
To confirm the occurrence of the production of the radical, the photocatalytic test at 1 g/L as catalyst loading and 9.4 W/m2 was repeated by adding 0.015 mL of a solution of 20.7 mM of tert-buryl alcohol as ·OH scavenger. The results are reported in Fig. 11 were evident the effect of TBA:
the removal of MB decreases drastically (from 57 to 20% in 3 h) with the addition of the scavenger.
This result confirms the crucial role of the hydroxyl radical in dye oxidation [8, 19, 21]. The MB removal in tests with TBA was comparable with the effect of the only adsorption mechanism on the catalyst surface [8, 21].
3.6 Kinetic studies
Removal kinetics were also studied, and pseudo-first-order kinetics (Eq. 6) was defined as a model that best describes the experimental results collected (Fig. 8b, Fig. 9b and Fig. 10b).
$$\text{ln}\left(\frac{{c}_{0}}{{c}_{t}}\right)=k*t$$
6
Where ct is the contaminant concentration at different times t (in min), c0 the contaminant concentration at time 0, and k the kinetic constant of reaction expressed in min− 1. The results of the kinetic constant, as slope of the linear data fitting where in the plots the time has been reported on the x-axis and \(\text{ln}\left(\frac{{C}_{0}}{{C}_{t}}\right)\)on the y-axis., are summarized in Table 3 and are in line with many studies from the literature [13], [20]– [24].
Table 3
Kinetic constant and R2 values for the photocatalytic degradation of MB at different MB concentrations.
MB concentration (ppm) | Catalyst loading (g/L) | Irradiance (W/m2) | k (min− 1) | R2 |
5 | 1 | 9.4 | 0.0036 | 0.9503 |
10 | 1 | 9.4 | 0.0036 | 0.9371 |
15 | 1 | 9.4 | 0.0032 | 0.8659 |
30 | 1 | 9.4 | 0.0017 | 0.8757 |
5 | 0.25 | 9.4 | 0.0016 | 0.9292 |
5 | 0.50 | 9.4 | 0.0017 | 0.8927 |
5 | 1 | 81.6 | 0.0087 | 0.9933 |
5 | 1 | 41.4 | 0.0051 | 0.9510 |
5 | 1 | 17.6 | 0.0041 | 0.9190 |
The kinetic constant revealed that the parameter that strongly influences the process is irradiation intensity. Although the values of the constants are lower than those reported in other works [25, 26], this effect is attributable to the low energy efficiency of the source used.