Morphologies and Crystal Structures
Figure 1 presents the strategy for preparing TiO
2/HNRs@SiO
2 from GNRs. First, it involves the hydrothermal synthesis of GNRs. Second, it uses a SiO
2 coating on the surface of GNRs to form GNRs@SiO
2. The SiO
2 coating serves dual purposes: to promote dispersion and prevent agglomeration and passivate the surface of the Fe-containing nanorod core to prevent
ilmenite (FeTiO3) formation
from the thermal annealing process, which lowers the photocatalytic performance. Third, it involves growing amorphous TiO
2 on the GNRs@SiO
2 surface and simultaneously transforming the GNRs to HNR under hydrothermal conditions to form TiO
2/HNRs@SiO
2. Finally, they are annealed thermally to form photocatalytically active TiO
2/HNRs@SiO
2.
Figure 2A–C presents the representative FESEM and inset TEM images of GNRs, GNRs@SiO
2, and TiO
2/HNRs@SiO
2 before thermal annealing, showing the unique morphologies of the SiO
2 coating on the GNRs and flower-like TiO
2 nanostructures grown on the SiO
2. The XRD patterns (
Figure 2D) of the GNRs and GNRs@SiO
2 revealed major diffraction peaks at 2θ = ~21.3, 33.3, 34.8, 36.7, and 53.3°, corresponding to the (110), (130), (021), (111), and (221) planes of α-FeOOH (JCPDS No. 29-0713) [
24], respectively, with an orthorhombic crystal structure because of the inner core of GNRs. The XRD patterns of TiO
2/HNRs@SiO
2 (
Figure 2D) showed peaks at 2θ = ~24.1, 33.1, 35.6, and 54.0°, which were indexed to the (012), (104), (110), and (116) planes of α-Fe
2O
3 (JCPDS No. 33-0664) with a rhombohedral crystal structure [
33]. Their peak intensities were less pronounced than those of α-Fe
2O
3 because of the inner core of TiO
2/HNRs@SiO
2, possibly due to the presence of an amorphous TiO
2 coating.
The FT-IR spectra of GNRs and GNRs@SiO
2 (
Figure 2E) exhibited the IR bands at 894 and 794 cm
–1, which were assigned to Fe–O vibrational modes for α-FeOOH [
24] before thermal annealing. These IR bands disappeared in the FT-IR spectrum of TiO
2/HNRs@SiO
2 (
Figure 2E), suggesting a complete thermal transformation of the inner core from α-FeOOH to α-Fe
2O
3. In addition, the IR bands at 1085 and 472 cm
–1 in the FT-IR spectra of GNRs@SiO
2 and TiO
2/HNRs@SiO
2, which were assigned to symmetric bending Si–O–Si vibrational modes in SiO
2, indicated the presence of SiO
2 coating on the nanorods [
34].
Figure 3 shows representative FESEM images and the corresponding XRD patterns of TiO
2/HNRs@SiO
2 thermally annealed at 400, 600, and 800 ℃. The XRD patterns of thermally annealed TiO
2/HNRs@SiO
2 (
Figure 3D) confirmed the major diffraction peaks for the anatase (JCPDS No. 21-1276) and rutile (JCPDS No. 21-1272) phases of TiO
2 [
35]. The relative fraction of the rutile and anatase TiO
2 phases in the thermally annealed TiO
2/HNRs@SiO
2 was determined quantitatively from the Rietveld refinement, respectively [
31,
36].
Figure S1 presents a typical result of the refinement, and
Table S1 lists the relative quantities of the phases modeled with the Rietveld refinement. All refinements consistently produced converged GoF values of less than 1.40, indicating excellent refinement quality [
37]. The relative anatase/rutile TiO
2 fraction in the TiO
2/HNRs@SiO
2 was 100/0, 81.4/18.6, and 6.3/93.7 (%/%) at annealing temperatures of 400, 600, and 800 ℃, respectively. Hence, the ratio between rutile and anatase TiO
2 increases as the annealing temperature increases.
The survey XPS spectra of TiO
2/HNRs@SiO
2 thermally annealed at 400, 600, and 800 ℃ over the full energy range confirmed the presence of Fe, Ti, O, and Si. (
Figure S2) High-resolution XPS spectra of TiO
2/HNRs@SiO
2 annealed at 400, 600, and 800 ℃ (
Figures S3 and S4) were deconvoluted to obtain their best fits and the following binding energies (BEs) summarized in
Table S2. Two peaks at approximately 723.2 and 710.2 eV with a peak separation of 13.0 eV (
Figure S3) from TiO
2/HNRs@SiO
2 annealed at 400, 600, and 800 ℃ were consistent with the doublet of Fe(III) 2p
1/2 and 2p
3/2 reported for hematite nanostructures [
33]. In addition, two peaks at approximately 464.2 and 458.5 eV with a peak separation of 5.7 eV (
Figure S3) from TiO
2/HNRs@SiO
2 annealed at 400 ℃ were in excellent agreement with the doublet of Ti(IV) 2p
1/2 and 2p
3/2 reported for nanostructured anatase TiO
2 [
38]. In addition, the shoulder peak at 460.6 eV from TiO
2/HNRs@SiO
2 annealed at 600 and 800 ℃ was assigned to Ti(III) 2p
1/2, which was in good agreement with Ti(III) because of the presence of Ti
2O
3 or oxygen vacancies [
39,
40]. The O 1s peak at 529.7 eV was similar to the BE of the lattice oxygen in anatase and rutile TiO
2 [
41]. The other O 1s peaks at 531.7 and 532.8 eV were assigned to the presence of Ti
2O
3 or non-lattice oxygen mixed with the Fe–O–Si/Ti–O–Si and Si–O–Si moieties, respectively [
42,
43,
44,
45]. The Si 2p peaks at 101.6 and 103.3 eV were assigned to the presence of silicon suboxides (i.e., Fe–O–Si or Ti–O–Si) and fully oxidized silicon (i.e., Si–O–Si), respectively [
42,
44,
46]. Our findings suggest that the samples become oxygen-rich and develop oxygen vacancies as the annealing temperature increases, leading to the partial conversion of Ti
4+ to Ti
3+, which is attributed to the transitioning of the anatase into the rutile phase [
47]. In addition, these results are consistent with those from the XRD patterns of TiO
2/HNRs@SiO
2 annealed at 400, 600, and 800 ℃.
Figure 4A presents UV-VIS DRS spectra of TiO
2/HNRs@SiO
2 annealed at 400, 600, and 800 ℃. Their optical spectra included strong absorption in the visible wavelength region. The optical bandgap (E
g) was estimated using the Tauc’s equation [
48]:
where
h,
ν,
B, and
n are Planck’s constant (4.1357 × 10
-15 eV·s), light frequency, a constant, and a coefficient associated with the electronic transition (i.e.,
n = 2 for indirect allowed transitions).
Figure 4B presents the plots of
(αhν)
2 versus
hν. The value of E
g, was determined by extrapolating the straight-line portion of plots of the energy axis (
Figure 4B). The E
g value TiO
2/HNRs@SiO
2 annealed at 400, 600, and 800 ℃ was 1.99, 1.99 and 2.03 eV, respectively. These results suggest that thermally activated TiO
2/HNRs@SiO
2 photocatalysts improve the light absorption characteristics in the visible region of the solar spectrum compared to pure TiO
2 systems, likely due to the formation of TiO
2/α-Fe
2O
3 heterostructures. α-Fe
2O
3 has a relatively narrow bandgap (i.e., ~1.9–2.2 eV) compared to traditional TiO
2-based photocatalysts with a wide bandgap (i.e., ~3.0–3.2 eV). Despite its favorable narrow bandgap, α-Fe
2O
3 suffers from rapid electron–hole recombination, which limits its photocatalytic efficiency. The narrow bandgap of thermally activated TiO
2/HNRs@SiO
2 indicates the formation of TiO
2/α-Fe
2O
3 heterostructures to improve light absorption and charge separation, highlighting their suitability and efficiency under natural sunlight.
Figure 5A and B present the photocatalytic MB degradation kinetics using thermally activated TiO
2/HNRs@SiO
2 as a function of the TiO
2/HNRs@SiO
2 dosage and MB concentration under ~150 min solar irradiation. First, the removal efficiency via photocatalytic decomposition using TiO
2/HNRs@SiO
2 annealed at 400 ℃ increased to ~83.8% as the TiO
2/HNRs@SiO
2 dosage was increased from 25 to 100 mg/mL under solar irradiation using a 10 mg/L MB concentration at pH = ~7 and 25 ℃. Second, the removal efficiency increased to ~95% as the MB concentration was decreased from 15 to 5 mg/L using 75 mg/L TiO
2/HNRs@SiO
2 annealed at 400 ℃ at pH = ~7 and 25 ℃. These results suggest that higher MB concentrations lead to more adsorbed MB molecules, leaving fewer active sites available and reducing the number of photons reaching the photocatalyst surface, which lowers the photoinduced generation of reactive oxygen species (ROS), such as hydroxyl, superoxide, and peroxide radicals [
3].
Figure 6A–C show the photocatalytic MB degradation kinetics of TiO
2/HNRs@SiO
2 activated by annealing at 400, 600, and 800 ℃ as a function of the solution pH = 4, 7, and 10, respectively. First, the removal efficiency of TiO
2/HNRs@SiO
2 annealed at 400 ℃ was ~43.8% at pH = 4.0. The removal efficiency decreased slightly to ~31.0% at pH = 4.0 as the annealing temperature was increased to 600 ℃. The removal efficiency of TiO
2/HNRs@SiO
2 annealed at 800 ℃ was ~28.8%, similar to that at 600 ℃. Second, the removal efficiencies of activated TiO
2/HNRs@SiO
2_400 increased up to ~89.2% as the solution pH was increased from 4.0 to 10.0. These results were comparable to or marginally higher than those reported for TiO
2 nanoparticles [
49], bismuth nanoparticles coated on silicon nanowires [
50], NiSe 2-ZnO nanocomposites [
51], and Ni-ZIF-8 metal-organic frameworks [
52], highlighting the effectiveness of TiO
2/HNRs@SiO
2 photocatalysts (
Table S3).
Photocatalytic MB degradation involves the adsorption of MB molecules onto the surface of activated TiO
2/HNRs@SiO
2. The photocatalytic experiments include 30 min to reach adsorption–desorption equilibrium in the dark, followed by irradiation of the light source to initiate the photoreaction. The improved removal efficiency at a pH > 7 was attributed to the enhanced interactions between the MB molecules and activated TiO
2/HNRs@SiO
2. The nature of the interactions was noncovalent and could be described either by physical or electrostatic adsorption. The physical adsorption of MB increases as the specific surface area of the activated TiO
2/HNRs@SiO
2 increases. Hence, the specific surface areas of activated TiO
2/HNRs@SiO
2 were determined from their low-temperature N
2 adsorption-desorption isotherms using the BET model.
Figure S5 presents the N
2 sorption isotherms of activated TiO
2/HNRs@SiO
2. The shape of the TiO
2/HNRs@SiO
2 isotherm annealed at 400 ℃ was similar to that of the characteristic type IV isotherm described in the IUPAC classification [
53]. In contrast, the TiO
2/HNRs@SiO
2 isotherms annealed at 600 and 800 ℃ were similar to those of the characteristic type II isotherm described in the IUPAC classification [
53]. The measured specific surface areas of TiO
2/HNRs@SiO
2 annealed at 400, 600, and 800 ℃ were 281.1, 89.7, and 36.8 m
2/g, respectively (
Table S4). Hence, the amount of MB adsorbed on the surface of activated TiO
2/HNRs@SiO
2 is proportional to its specific surface area.
Figure S6 shows the zeta potentials of TiO
2/HNRs@SiO
2 annealed at 400, 600, and 800 ℃. Considering the pK
a of MB (pK
a = ~3.8), the predominant MB species are cationic in aqueous solutions at pH > 3.8 [
54]. The zeta potential of TiO
2/HNRs@SiO
2 annealed at 400, 600, and 800 ℃ was –29.4 ± 0.6, –24.7 ± 0.5 and –30.7 ± 0.7 mV at pH = ~7.0, respectively. Therefore, higher MB adsorption occurred on the surface of negatively charged, activated TiO
2/HNRs@SiO
2. Furthermore, the superior MB removal efficiency of TiO
2/HNRs@SiO
2 annealed at 400 ℃ compared to 600 ℃ and 800 ℃ was attributed to the higher adsorption, driven by the synergistic effects of increased specific surface area and enhanced electrostatic interactions (e.g., higher negative zeta potentials) at neutral and higher pH. The MB molecules must be in close proximity to the active sites of the photocatalytic TiO
2/HNRs@SiO
2 in order for the photoreaction to occur. Accordingly, higher adsorption ensures that more MB molecules are localized on the photocatalyst surface, allowing direct interactions with the photochemically induced ROS, providing the rationale for enhanced removal efficiency.
The adsorption–desorption equilibrium of MB molecules on the active sites of the photocatalytic TiO
2/HNRs@SiO
2 under dark conditions dictates the initial amount of MB molecules that will be photodegraded when the photoirradiation starts. Electron–hole pairs are formed when the photocatalysts are exposed to light with an energy equal to or greater than their band gap. Electrons and holes, if not recombined, participate in surface reactions to generate ROS that attack and break down the adsorbed MBs into smaller intermediates and ultimately to CO
2, H
2O, and other non-toxic compounds. Three assumptions were proposed to establish adsorption equilibrium, a rate-determining step for the photocatalytic MB degradation kinetics. First, the rate of MB desorption is relatively insignificant and can be neglected, leading to adsorption as an irreversible process. Second, the MB concentration in solution is relatively constant during adsorption. Third, the photodegradation, followed by the desorption of photochemically decomposed intermediates, is relatively faster than the adsorption. These assumptions suggest that the total removal efficiency can be estimated as the sum of two factors: 1) the adsorption of MB under dark conditions (i.e., dark efficiency) and 2) the photochemical decomposition of the adsorbed MB when exposed to light (i.e., light efficiency).
Table S5 lists these efficiencies. For example, TiO
2/HNRs@SiO
2 annealed at 400 ℃ exhibited an increase in removal efficiency from 33.8% in the dark to 75.4% after exposure to light. TiO
2/HNRs@SiO
2 annealed at 600 and 800 ℃ showed much lower dark efficiencies of 5.6 and 7.7%, which improved to 51.4 and 55.1% under light, respectively. The net light efficiency of TiO
2/HNRs@SiO
2 annealed at 400, 600, and 800 ℃ from photochemical decomposition was 41.6, 45.8, and 47.4%, respectively. These results suggest that the photochemical decomposition efficiency increased slightly as the annealing temperature increased, likely because of the rutile-to-anatase TiO
2 ratio of thermally annealed TiO
2/HNRs@SiO
2. In particular, with its higher rutile content, TiO
2/HNRs@SiO
2 annealed at 800 ℃ exhibited enhanced light efficiency because the visible light absorption and charge transfer characteristics improved at a higher rutile-to-anatase ratio [
55]. These findings are consistent with previous studies highlighting the significant impact of the anatase-to-rutile mixed ratio on photocatalytic performance [
55,
56,
57].
Figure S7 presents the results of scavenger tests to identify and evaluate the contributions of different reactive species involved in the photocatalytic degradation of MB. Scavengers such as t-BuOH and EDTA react selectively with particular reactive species, such as hydroxyl radicals (HO·) and photogenerated holes (
h+), respectively. The removal efficiency of MB by TiO
2/HNRs@SiO
2 annealed at 400 ℃ decreased by 6.7 and 31.7% after adding t-BuOH and EDTA, respectively, compared to the control without scavengers. EDTA captures
h+, reducing its availability for the direct oxidation of MB and ·OH generation by
h+ or water oxidation, suggesting that
h+ plays a significant role in breaking down MB during photocatalysis [
58].
Figure 7 outlines the potential mechanism for photocatalytic MB decomposition using activated TiO
2/HNRs@SiO
2 nanocomposite photocatalysts. An electron–hole pair is produced when TiO
2/HNRs@SiO
2 nanocomposite photocatalysts are irradiated with visible light. The TiO
2/α-Fe
2O
3 heterostructures within the nanorods facilitate the charge separation, effectively reducing their recombination. The resulting photogenerated electrons (
e–) can react with dissolved oxygen to form superoxide radicals (O
2·–), while the photogenerated holes (
h+) react with water or hydroxide ions (H
2O/OH
–) to produce hydroxyl radicals (HO·). These can be pathways to generate reactive oxygen species that can promote the efficient breakdown of MB dyes.