Non-Stacked γ-Fe2O3/C@TiO2 Double-Layer Hollow Nanoparticles for Enhanced Photocatalytic Applications under Visible Light

Herein, a non-stacked γ-Fe2O3/C@TiO2 double-layer hollow nano photocatalyst has been developed with ultrathin nanosheets-assembled double shells for photodegradation phenol. High catalytic performance was found that the phenol could be completely degraded in 135 min under visible light, due to the moderate band edge position (VB at 0.59 eV and CB at −0.66 eV) of the non-stacked γ-Fe2O3/C@TiO2, which can expand the excitation wavelength range into the visible light region and produce a high concentration of free radicals (such as ·OH, ·O2−, holes). Furthermore, the interior of the hollow composite γ-Fe2O3 is responsible for charge generation, and the carbon matrix facilitates charge transfer to the external TiO2 shell. This overlap improved the selection/utilization efficiency, while the unique non-stacked double-layered structure inhibited initial charge recombination over the photocatalysts. This work provides new approaches for photocatalytic applications with γ-Fe2O3/C-based materials.


Introduction
Phenol containing wastewater is produced in many industries, such as pharmaceuticals, polymer, dye, etc. [1]. Once phenol is released into the environment, it cannot be degraded inherently under the natural environment, and it causes harm to human health [2,3]. The phenol degradation is complex due to the high conjugated molecular system [4] and lower phenol concentration is also toxic [2]. Therefore, designing an environmentally friendly, cost-effective and highly efficient approach for phenol treatment is urgent. Photocatalysis is a simple technology among water treatment technologies [5][6][7], owing to its high mineralization and sturdy treatment efficiency [8,9].
Titanium dioxide (TiO 2 ) is the most-used semiconductor material in the photocatalytic field, due to its chemical/physical excellent profound mineralization capability and stability under ultraviolet light [10]. However, two significant problems need to be solved: The low specific surface area of TiO 2 (50 m 2 /g for P25, TiO 2 with a mixed rutile phase, and anatase phase with an average particle size of 25 nm) confines its adsorption capacity for pollutant molecules. Additionally, it has a wide energy gap and is challenging to utilize visible light. Therefore, the preparation of a TiO 2 catalyst with an extensive specific surface area and response to visible light is the bottleneck problem in applying TiO 2 .
2.1.2. Synthesis of Non-Stacked γ-Fe 2 O 3 /C@TiO 2 Double-Layer Hollow Nanoparticles Figure 1 shows the synthesis approaches for the non-stacked γ-Fe 2 O 3 /C@TiO 2 doublelayer hollow nanoparticles. In this fraction, the nanoparticles were prepared by the following steps.
(1) Synthesis of SiO 2 @γ-Fe 2 O 3 /C nanoparticles  Figure 1 shows the synthesis approaches for the non-stacked γ-Fe2O3/C@TiO2 d layer hollow nanoparticles. In this fraction, the nanoparticles were prepared by t lowing steps. (1) Synthesis of SiO2@γ-Fe2O3/C nanoparticles Monodispersed SiO2 nanoparticles were first synthesized through the method. A modified hydrothermal approach prepared the core-shell SiO2/γ-Fe2O3 noparticles. Silica nanoparticles (100 mg) and ferrocene (200 mg) were suspended anol (65 mL) using ultrasonic for 30 min. Then, 2 mL H2O2 was dropwise counted in mixture and cruised with energetic stirring for 1.5 h. A Teflon-lined stainless-stee clave(TEFIC BIOTECH CO., Xi'an, China) was used to heat the homogeneous solu 210 °C for 48 h. The product was obtained after the autoclave cooled down to room perature (3 h). Acetone and ethanol were used to wash the obtained products 3 respectively, and then the product was vacuum-dried for 12 h.
(2) Synthesis of SiO2@γ-Fe2O3/C@SiO2 nanoparticles The Fe2O3/C@SiO2 nanoparticles (0.3 g) obtained from the last step were adde a three-neck round-bottom flask filled with the ammonia solution (5.0 mL, 28 wt%) nol anhydrous (280 mL), and deionized water (70 mL), ultrasonicate for 20 min. TE mL) was added dropwise to the mixture in 10 min. The reaction mixture was mechan stirred for 10 h at room temperature before the products separated. Then the final pr was then washed 3 times, respectively, using ethanol and deionized water, and dr vacuum.
(3) Synthesis of SiO2@γ-Fe2O3/C@SiO2@TiO2 nanoparticles The core-shell SiO2@γ-Fe2O3/C@SiO2@TiO2 nanoparticles were synthesized usi condensation and hydrolysis of TBOT. Ethanol anhydrous (200 mL) was used to di the obtained SiO2@γ-Fe2O3/C@SiO2 nanoparticles (0.15 g) from the above procedure monia solution (0.9 mL, 28 wt%) was counted to the mixture and ultrasonicated min. TBOT (2.0 mL) was then added dropwise to the mixture in 5 min, followed b tinuous mechanical stirring for 24 h at 45 °C. The products were separated by centr tion and washed 3 times, using ethanol and deionized water, and dried under vacu (4) Synthesis of non-stacked γ-Fe2O3/C@TiO2 double-layer hollow nanoparticles An alkaline hydrothermal etching-assisted crystallization approach was used t thesize the final products, the SiO2@γ-Fe2O3/C@SiO2@TiO2 nanoparticles. The Si Fe2O3/C@SiO2@TiO2 nanoparticles (0.4 g) obtained above were added into a Teflon stainless-steel autoclave (50 mL) filled with a NaOH aqueous solution (25 mL, 2.0 M autoclave was sealed and heated to 100 °C for 4 h and cooled to room temperature before the next step. The final SiO2@γ-Fe2O3/C@SiO2@TiO2 nanoparticles were subm in a HCl aqueous solution (100 mL, 0.1 M) for 20 min followed by washing with deio Monodispersed SiO 2 nanoparticles were first synthesized through the Stöber method. A modified hydrothermal approach prepared the core-shell SiO 2 /γ-Fe 2 O 3 /C nanoparticles. Silica nanoparticles (100 mg) and ferrocene (200 mg) were suspended in ethanol (65 mL) using ultrasonic for 30 min. Then, 2 mL H 2 O 2 was dropwise counted into the mixture and cruised with energetic stirring for 1.5 h. A Teflon-lined stainless-steel autoclave(TEFIC BIOTECH CO., Xi'an, China) was used to heat the homogeneous solution at 210 • C for 48 h. The product was obtained after the autoclave cooled down to room temperature (3 h). Acetone and ethanol were used to wash the obtained products 3 times, respectively, and then the product was vacuum-dried for 12 h.
(2) Synthesis of SiO 2 @γ-Fe 2 O 3 /C@SiO 2 nanoparticles The Fe 2 O 3 /C@SiO 2 nanoparticles (0.3 g) obtained from the last step were added into a three-neck round-bottom flask filled with the ammonia solution (5.0 mL, 28 wt%), ethanol anhydrous (280 mL), and deionized water (70 mL), ultrasonicate for 20 min. TEOS (4 mL) was added dropwise to the mixture in 10 min. The reaction mixture was mechanically stirred for 10 h at room temperature before the products separated. Then the final product was then washed 3 times, respectively, using ethanol and deionized water, and dried in vacuum.
(3) Synthesis of SiO 2 @γ-Fe 2 O 3 /C@SiO 2 @TiO 2 nanoparticles The core-shell SiO 2 @γ-Fe 2 O 3 /C@SiO 2 @TiO 2 nanoparticles were synthesized using the condensation and hydrolysis of TBOT. Ethanol anhydrous (200 mL) was used to disperse the obtained SiO 2 @γ-Fe 2 O 3 /C@SiO 2 nanoparticles (0.15 g) from the above procedure. Ammonia solution (0.9 mL, 28 wt%) was counted to the mixture and ultrasonicated for 15 min. TBOT (2.0 mL) was then added dropwise to the mixture in 5 min, followed by continuous mechanical stirring for 24 h at 45 • C. The products were separated by centrifugation and washed 3 times, using ethanol and deionized water, and dried under vacuum.
(4) Synthesis of non-stacked γ-Fe 2 O 3 /C@TiO 2 double-layer hollow nanoparticles An alkaline hydrothermal etching-assisted crystallization approach was used to synthesize the final products, the SiO 2 @γ-Fe 2 O 3 /C@SiO 2 @TiO 2 nanoparticles. The SiO 2 @γ-Fe 2 O 3 /C@SiO 2 @TiO 2 nanoparticles (0.4 g) obtained above were added into a Teflon-lined stainless-steel autoclave (50 mL) filled with a NaOH aqueous solution (25 mL, 2.0 M). The autoclave was sealed and heated to 100 • C for 4 h and cooled to room temperature in 3 h before the next step. The final SiO 2 @γ-Fe 2 O 3 /C@SiO 2 @TiO 2 nanoparticles were submerged in a HCl aqueous solution (100 mL, 0.1 M) for 20 min followed by washing with deionized water until pH value was close to 7. The final nanoparticles were then dried at 60 • C for 5 h. Finally, the catalyst was annealed at 400 • C for 4 h in an oxygen-deficient environment.

Catalyst Characterization
X-ray powder diffraction (XRD) patterns with 2θ ranging from 10 to 80 • (40 kV and 30 mA, D/MAX-2500, Rigaku, Japan), as well as transmission electron microscopy (TEM, JEM-2100 Felectron microscope operating at 200 kV, Tokyo, Japan) were used to obtain the crystal and morphological structure. The samples' specific surface areas were characterized by the Brunauer-Emmett-Teller (BET) model. Quadrasorb SI analyzer was used to obtain the N 2 adsorption-desorption isotherms at 77 K. The Barrett-Joyner-Halenda (BJH) model was used to analyze the pore size and volume. X-ray photoelectron spectroscopy (XPS) was performed using a 5300 ESCA equipment (PerkinElmer PHI Co.,Hopkinton, Massachusetts, USA) with an Al Kα X-ray source (250 W) to investigate the chemical compositions of samples. A UV-vis spectrophotometer (UV-2600, Shimadzu, Tokyo, Japan) with BaSO 4 was used to acquire the UV-vis diffuse reflection spectra (UV-vis DRS).

Catalytic Activity Measurement
The photocatalytic activity was evaluated using a photochemical reactor (TG-10B, Beijing, China) under xenon lamp (300 W). In each experiment, 0.1 g catalyst was weighed in a quartz glass tube, and 30 mL phenol (20 mg L −1 ) was added into it. Before photocatalytic reaction, a dark reaction was carried out for 30 min to achieve the equilibrium of adsorption and desorption of phenol. The absorbency of phenol was determined by a UV-VIS spectrophotometer.
According to absorbance conversion of phenol, the formula of photocatalytic reaction removal rate was shown in (1).
A 0 -the initial absorbance of phenol. A t -Absorption of phenol at t min.

Textural Properties of Catalysts
The wide-angle XRD pattern of non-stacked γ-Fe 2 O 3 /C@TiO 2 is shown in Figure 2a. The characteristic peaks of the catalyst were indexed to anatase TiO 2 (JCPDS-ICDD21-1272) [21][22][23]. Other phases such as γ-Fe 2 O 3 were not observed because the thick TiO 2 shell passivated the X-ray diffraction. To better analyse the phase composition of the catalyst inner layer, the precursor (hollow γ-Fe 2 O 3 /C) of γ-Fe 2 O 3 /C@TiO 2 was characterized by XRD also. As shown in Figure 2b, the diffraction peaks of γ-Fe 2 O 3 (JCPDS-ICDD 25-1402) were observed in hollow γ-Fe 2 O 3 /C [24,25]. Figure 2c shows the Raman spectra of hollow γ-Fe 2 O 3 /C, stacked γ-Fe 2 O 3 /C@TiO 2 , and non-stacked γ-Fe 2 O 3 /C@TiO 2 . The five vibrational modes, located at 640, 545, 395, 193, and 142 cm −1 , represent E g Raman active, A 1g + B 1g , B 1g , e.g., and E g modes, indicating that the anatase is the dominant phase of the TiO 2 hollow spheres. The N2 adsorption-desorption was used to measure the specific surface area of the catalyst. Type IV isothermal curves and N2 hysteresis loops were observed in Figure 3a, which indicated that the non-stacked γ-Fe2O3/C@TiO2 was a mesoporous material. The pore volume and pore size distributions were also characterized. As shown in Figure 3b, the pore size was mainly distributed from 5-30 nm, which indicated the catalysts' mesoporous nature. The hollow structure expected to endow the material with a larger specific surface area (145.328 m 2 /g), which has a strong ability to absorb pollution molecules so that the active species produced on its surface can directly react with pollutants and en- The N 2 adsorption-desorption was used to measure the specific surface area of the catalyst. Type IV isothermal curves and N 2 hysteresis loops were observed in Figure 3a, which indicated that the non-stacked γ-Fe 2 O 3 /C@TiO 2 was a mesoporous material. The pore volume and pore size distributions were also characterized. As shown in Figure 3b, the pore size was mainly distributed from 5-30 nm, which indicated the catalysts' mesoporous nature. The hollow structure expected to endow the material with a larger specific surface area (145.328 m 2 /g), which has a strong ability to absorb pollution molecules so that the active species produced on its surface can directly react with pollutants and enhance the catalytic performance. Figure 2. XRD patterns of non-stacked γ-Fe2O3/C@TiO2 (a) and hollow γ-Fe2O3/C (b); Raman spe (c) of non-stacked γ-Fe2O3/C@TiO2, stacked γ-Fe2O3/C@TiO2, and hollow γ-Fe2O3/C. The N2 adsorption-desorption was used to measure the specific surface area of catalyst. Type IV isothermal curves and N2 hysteresis loops were observed in Figure  which indicated that the non-stacked γ-Fe2O3/C@TiO2 was a mesoporous material. T pore volume and pore size distributions were also characterized. As shown in Figure the pore size was mainly distributed from 5-30 nm, which indicated the catalysts' me porous nature. The hollow structure expected to endow the material with a larger spec surface area (145.328 m 2 /g), which has a strong ability to absorb pollution molecules that the active species produced on its surface can directly react with pollutants and hance the catalytic performance.  Figure 4a', the γ-Fe2O3/C layer was firmly attached to the outer surf of the SiO2 sphere, and the wrapped SiO2@γ-Fe2O3/C nanoparticles exhibited relativ uniform spherical structures. Then, the thin layers of SiO2 and TiO2 were successfu coated on the surface of the SiO2@γ-Fe2O3/C core to form uniform SiO2@ Fe2O3/C@SiO2@TiO2 nanoparticles (Figure 4b'). After treatment in a hot alkaline soluti the silicon component was removed, and the TiO2 expanded both inside and outside form non-stacked γ-Fe2O3/C@TiO2 (Figure 4c'). ICP-AES results showed the nominal ra of the components (Fe:Ti) was 3.9:1. Furthermore, from Figure 4a,b, almost all catal particles can maintain this particular structure. A clear lattice plane was exposed w interplanar spacings of 0.29 nm and 0.35 nm, which revealed the presence of anatase T and γ-Fe2O3/C, respectively, which mainly expressed the (220) and (101)   As shown in Figure 4a', the γ-Fe 2 O 3 /C layer was firmly attached to the outer surface of the SiO 2 sphere, and the wrapped SiO 2 @γ-Fe 2 O 3 /C nanoparticles exhibited relatively uniform spherical structures. Then, the thin layers of SiO 2 and TiO 2 were successfully coated on the surface of the SiO 2 @γ-Fe 2 O 3 /C core to form uniform SiO 2 @γ-Fe 2 O 3 /C@SiO 2 @TiO 2 nanoparticles (Figure 4b'). After treatment in a hot alkaline solution, the silicon component was removed, and the TiO 2 expanded both inside and outside to form non-stacked γ-Fe 2 O 3 /C@TiO 2 (Figure 4c'). ICP-AES results showed the nominal ratio of the components (Fe:Ti) was 3.9:1. Furthermore, from Figure 4a,b, almost all catalyst particles can maintain this particular structure. A clear lattice plane was exposed with interplanar spacings of 0.29 nm and 0.35 nm, which revealed the presence of anatase TiO 2 and γ-Fe 2 O 3 /C, respectively, which mainly expressed the (220) and (101)   To further investigate the composition distribution of the inner shell, the Fe, C, and O species located in hollow γ-Fe2O3/C were characterized by XPS. The typical peaks of Fe2p3/2 and Fe2p1/2 were located at 711.0 and 724.5 eV, respectively (Figure 5a) [26][27][28]. The presence of the characteristic satellite peak at 719.0 eV confirmed the formation of γ-Fe2O3, which is the key characteristic distinguishing Fe2O3 from Fe3O4. In the O1s spectra ( Figure  5b), the presence of Fe-O-C indicated strong interactions between γ-Fe2O3 and C, which may serve as an electron pathway to facilitate electron transfer via Fe-O-C. Two small peaks at 285.9 and 288.7 eV correspond to the residual C = O and C-O species in the nanoshell after calcination (Figure 5c). The distribution of Ti and O species located in the shell of non-stacked γ-Fe2O3/C@TiO2 is shown in Figure 5d,e. The peaks at binding energies of 464.9 and 458.9 eV were assigned to the Ti2p1/2 and Ti2p3/2 core levels of Ti 4+ , respectively. The two peaks located at 458.2 and 463.7 eV corresponded to the characteristic peaks of Two small peaks at 285.9 and 288.7 eV correspond to the residual C = O and C-O species in the nano-shell after calcination (Figure 5c). The distribution of Ti and O species located in the shell of non-stacked γ-Fe 2 O 3 /C@TiO 2 is shown in Figure 5d,e. The peaks at binding energies of 464.9 and 458.9 eV were assigned to the Ti2p 1/2 and Ti2p 3/2 core levels of Ti 4+ , respectively. The two peaks located at 458.2 and 463.7 eV corresponded to the characteristic peaks of Ti2p 3/2 and Ti2p 1/2 of Ti 3+ . The O1s spectra displayed two major oxygen peaks at 530.1 and 531.7 eV, which were attributed to lattice oxygen (O lat ) and surface-absorbed oxygen (O sur ), respectively. The lattice oxygen species are nucleophilic reagents that are usually responsible for oxidation reactions.

As shown in
To further investigate the composition distribution of the inner shell, the Fe, C, and O species located in hollow γ-Fe2O3/C were characterized by XPS. The typical peaks of Fe2p3/2 and Fe2p1/2 were located at 711.0 and 724.5 eV, respectively (Figure 5a) [26][27][28]. The presence of the characteristic satellite peak at 719.0 eV confirmed the formation of γ-Fe2O3, which is the key characteristic distinguishing Fe2O3 from Fe3O4. In the O1s spectra ( Figure  5b), the presence of Fe-O-C indicated strong interactions between γ-Fe2O3 and C, which may serve as an electron pathway to facilitate electron transfer via Fe-O-C. Two small peaks at 285.9 and 288.7 eV correspond to the residual C = O and C-O species in the nanoshell after calcination (Figure 5c). The distribution of Ti and O species located in the shell of non-stacked γ-Fe2O3/C@TiO2 is shown in Figure 5d,e. The peaks at binding energies of 464.9 and 458.9 eV were assigned to the Ti2p1/2 and Ti2p3/2 core levels of Ti 4+ , respectively. The two peaks located at 458.2 and 463.7 eV corresponded to the characteristic peaks of Ti2p3/2 and Ti2p1/2 of Ti 3+ . The O1s spectra displayed two major oxygen peaks at 530.1 and 531.7 eV, which were attributed to lattice oxygen (Olat) and surface-absorbed oxygen (Osur), respectively. The lattice oxygen species are nucleophilic reagents that are usually responsible for oxidation reactions.

UV-Vis Absorbance Spectra of Non-Stacked γ-Fe 2 O 3 /C@TiO 2
The optical absorption of TiO 2 was affected by impurities and changes in the bandgap. The impact of catalysts' overlap modes on the absorption of visible light was analyzed using UV-vis-DRS. Figure 6a shows that the three catalysts generated the absorption bandedge around 550 nm, corresponding to the transition of electrons from the top of the valence band to the footing of the conduction band. The absorption band edge did not immediately decline to 0 from 550 to 800 nm; however, it entered an extended buffer period, corresponding to the energy of photon needed to transition electrons from the O2p level to the impurity station. This extensive absorption range reflected the main characteristics of γ-Fe 2 O 3 /C.

UV-Vis Absorbance Spectra of Non-Stacked γ-Fe2O3/C@TiO2
The optical absorption of TiO2 was affected by impurities and changes in the bandgap. The impact of catalysts' overlap modes on the absorption of visible light was analyzed using UV-vis-DRS. Figure 6a shows that the three catalysts generated the absorption band-edge around 550 nm, corresponding to the transition of electrons from the top of the valence band to the footing of the conduction band. The absorption band edge did not immediately decline to 0 from 550 to 800 nm; however, it entered an extended buffer period, corresponding to the energy of photon needed to transition electrons from the O2p level to the impurity station. This extensive absorption range reflected the main characteristics of γ-Fe2O3/C.  Figure 6b, the first maximum at 2.3 eV (540 nm) is associated with point lattice defects, namely oxygen vacancies, while the subsequent growth at energies above 2.3 eV corresponds to the fundamental bandgap [29,30]. The corresponding bandgap energies of non-stacked γ-Fe2O3/C@TiO2 and stacked γ-Fe2O3/C@TiO2 were 1.25 and 1.32 eV (Figure  6b), which were both higher than hollow γ-Fe2O3/C (1.11 eV). The band locations ( Figure   Figure 6. UV-Vis diffuse reflectance spectra (a) and calculated band gap patterns based on UV-Vis diffuse reflectance spectra (b) of non-stacked γ-Fe 2 O 3 /C@TiO 2 , stacked γ-Fe 2 O 3 /C@TiO 2 , and hollow γ-Fe 2 O 3 /C. From Figure 6b, the first maximum at 2.3 eV (540 nm) is associated with point lattice defects, namely oxygen vacancies, while the subsequent growth at energies above 2.3 eV corresponds to the fundamental bandgap [29,30]. The corresponding bandgap energies of non-stacked γ-Fe 2 O 3 /C@TiO 2 and stacked γ-Fe 2 O 3 /C@TiO 2 were 1.25 and 1.32 eV (Figure 6b), which were both higher than hollow γ-Fe 2 O 3 /C (1.11 eV). The band locations (Figure 7) were calculated using the XPS valence spectra and bandgap energies. The γ- The charge generation and charge transfer behaviors over non-stacked γ-Fe2O3/C@TiO2 are proposed in Scheme 1. Incident light can induce the transition of γ-Fe2O3. Excited electrons generated in γ-Fe2O3 can be efficiently transmitted to the conduction band of TiO2 through C species while holes remain in the valence band of γ-Fe2O3, resulting in effective electron-hole separation. Therefore, the bandgap will become smaller, and the excitation wavelength range of the TiO2 composite was expanded into the visible light region. From Scheme 1, the ·OH, ·O 2− and holes will contribute to the degradation of organic matter, and the hydroxyl radicals are not obtained from hole oxidation of water but from the interaction of electrons, hydrogen ions, and superoxide radicals. And the production process of hydroxyl radical is as follows [11]: H2O2 + e − = ·OH + OH − Excited electrons generated in γ-Fe 2 O 3 can be efficiently transmitted to the conduction band of TiO 2 through C species while holes remain in the valence band of γ-Fe 2 O 3 , resulting in effective electron-hole separation. Therefore, the bandgap will become smaller, and the excitation wavelength range of the TiO 2 composite was expanded into the visible light region. From Scheme 1, the ·OH, ·O 2− and holes will contribute to the degradation of organic matter, and the hydroxyl radicals are not obtained from hole oxidation of water but from the interaction of electrons, hydrogen ions, and superoxide radicals. And the production process of hydroxyl radical is as follows [11]:

Photocatalytic Degradation of Phenol
Hollow γ-Fe 2 O 3 /C, stacked γ-Fe 2 O 3 /C@TiO 2 , and non-stacked γ-Fe 2 O 3 /C@TiO 2 photocatalysts were used to degrade phenol under visible light irradiation. As shown in Figure 8a, non-stacked γ-Fe 2 O 3 /C@TiO 2 showed the highest performance and thoroughly degraded phenol after 135 min of irradiation under Xe lamp with 300 W power (Figure 8b). The catalytic performance was as follows: non-stacked γ-Fe 2 O 3 /C@TiO 2 > stacked γ-Fe 2 O 3 /C@TiO 2 > hollow γ-Fe 2 O 3 /C. Compared with stacked γ-Fe 2 O 3 /C@TiO 2 , the performance of non-stacked γ-Fe 2 O 3 /C@TiO 2 was significantly enhanced due to the unique designed morphology and structure, which can enhance the absorption capacity of visible light and improve the activity of adsorption sites. The hollow inner cavity also permitted the scattering and refraction of light; therefore, the diffusion distance of electronhole pairs generated by light was reduced, maximizing light utilization. Simultaneously, the unique structure also expanded the specific surface area of the catalytic material. The carbon species can facilitate charge transfer from γ-Fe 2 O 3 to the outer TiO 2 , reducing the recombination probability of photogenerated carriers. The reusability of non-stacked γ-Fe 2 O 3 /C@TiO 2 was investigated ( Figure S1), and the catalytic activity of the catalyst decreased by only about 5.0% after 5 catalytic cycles.

Photocatalytic Degradation of Phenol
Hollow γ-Fe2O3/C, stacked γ-Fe2O3/C@TiO2, and non-stacked γ-Fe2O3/C@TiO2 ph catalysts were used to degrade phenol under visible light irradiation. As shown in Fi 8a, non-stacked γ-Fe2O3/C@TiO2 showed the highest performance and thoroughly graded phenol after 135 min of irradiation under Xe lamp with 300 W power (Figure The catalytic performance was as follows: non-stacked γ-Fe2O3/C@TiO2 > stacke Fe2O3/C@TiO2 > hollow γ-Fe2O3/C. Compared with stacked γ-Fe2O3/C@TiO2, the pe mance of non-stacked γ-Fe2O3/C@TiO2 was significantly enhanced due to the uniqu signed morphology and structure, which can enhance the absorption capacity of vi light and improve the activity of adsorption sites. The hollow inner cavity also perm the scattering and refraction of light; therefore, the diffusion distance of electron pairs generated by light was reduced, maximizing light utilization. Simultaneously unique structure also expanded the specific surface area of the catalytic material. The bon species can facilitate charge transfer from γ-Fe2O3 to the outer TiO2, reducing th combination probability of photogenerated carriers. The reusability of non-stacke Fe2O3/C@TiO2 was investigated ( Figure S1), and the catalytic activity of the catalys creased by only about 5.0% after 5 catalytic cycles. The degradation of phenol using the non-stacked γ-Fe2O3/C@TiO2 was analyse UV-Vis absorbance spectra (Figure 9a) and in situ DRIFTS (Figure 9b). Interestingly ure 9a shows that the absorption band (269 nm) of phenol over non-stacke Fe2O3/C@TiO2 substantially increased upon prolonging the irradiation time beyon min, suggesting the formation of more intermediate products. When the irradiation was prolonged, the intermediate products formed during the degradation of pheno composed to form small molecules, consistent with the results of in situ DRIFTS (F 9b). The absorption bands centered at 1410 cm −1 represented the stretching vibration OH functional group of phenol. The characteristic peaks of phenol significantly decre The degradation of phenol using the non-stacked γ-Fe 2 O 3 /C@TiO 2 was analysed by UV-Vis absorbance spectra (Figure 9a) and in situ DRIFTS (Figure 9b). Interestingly, Figure 9a shows that the absorption band (269 nm) of phenol over non-stacked γ-Fe 2 O 3 / C@TiO 2 substantially increased upon prolonging the irradiation time beyond 30 min, suggesting the formation of more intermediate products. When the irradiation time was prolonged, the intermediate products formed during the degradation of phenol decomposed to form small molecules, consistent with the results of in situ DRIFTS (Figure 9b). The absorption bands centered at 1410 cm −1 represented the stretching vibration of -OH functional group of phenol. The characteristic peaks of phenol significantly decreased when exposed to light after 30 min. This shows that the phenolic -OH was first oxidized during phenol degradation. To further investigate the degradation of phenol over non-stacked γ-Fe2O3/C@ the intermediate products (p-dihydroxybenzene, o-dihydroxybenzene, p-benzoqui oxalic acid, acetic acid, and formic acid) were measured by high-performance liquid matography (Figure 10a,b). Under light irradiation, the holes formed by γ-Fe2O3/C@ reacted with phenol to generate unstable free radicals containing phenoxy groups. T and p-phenolic hydroxyl groups tend to form stable o-dihydroxybenzene, pdroxybenzene, which were further oxidized to p-benzoquinone. Small molecules, su oxalic acid, acetic acid, and formic acid, were formed through ring-opening and eventually mineralized into CO2 and H2O. Non-stacked γ-Fe2O3/C@TiO2 had a mod band edge position, allowing it to excite hydroxyl radicals, superoxide free radicals holes to oxidize organic compounds. This was ascribed to an appropriate overlap m between the two layers. Moreover, Table 1 shows a comparison of the double-layer h nanoparticles with the other catalysts; it is found that the catalysts in this paper have b catalytic performance and can use fewer catalysts in a short time to achieve higher lytic efficiency.   To further investigate the degradation of phenol over non-stacked γ-Fe 2 O 3 /C@TiO 2 , the intermediate products (p-dihydroxybenzene, o-dihydroxybenzene, p-benzoquinone, oxalic acid, acetic acid, and formic acid) were measured by high-performance liquid chromatography (Figure 10a,b). Under light irradiation, the holes formed by γ-Fe 2 O 3 /C@TiO 2 reacted with phenol to generate unstable free radicals containing phenoxy groups. The o-and p-phenolic hydroxyl groups tend to form stable o-dihydroxybenzene, p-dihydroxybenzene, which were further oxidized to p-benzoquinone. Small molecules, such as oxalic acid, acetic acid, and formic acid, were formed through ring-opening and were eventually mineralized into CO 2 and H 2 O. Non-stacked γ-Fe 2 O 3 /C@TiO 2 had a moderate band edge position, allowing it to excite hydroxyl radicals, superoxide free radicals, and holes to oxidize organic compounds. This was ascribed to an appropriate overlap mode between the two layers. Moreover, Table 1 shows a comparison of the double-layer hollow nanoparticles with the other catalysts; it is found that the catalysts in this paper have better catalytic performance and can use fewer catalysts in a short time to achieve higher catalytic efficiency. To further investigate the degradation of phenol over non-stacked γ-Fe2O3/C@ the intermediate products (p-dihydroxybenzene, o-dihydroxybenzene, p-benzoqui oxalic acid, acetic acid, and formic acid) were measured by high-performance liquid matography (Figure 10a,b). Under light irradiation, the holes formed by γ-Fe2O3/C@ reacted with phenol to generate unstable free radicals containing phenoxy groups. T and p-phenolic hydroxyl groups tend to form stable o-dihydroxybenzene, pdroxybenzene, which were further oxidized to p-benzoquinone. Small molecules, su oxalic acid, acetic acid, and formic acid, were formed through ring-opening and eventually mineralized into CO2 and H2O. Non-stacked γ-Fe2O3/C@TiO2 had a mod band edge position, allowing it to excite hydroxyl radicals, superoxide free radicals holes to oxidize organic compounds. This was ascribed to an appropriate overlap between the two layers. Moreover, Table 1 shows a comparison of the double-layer h nanoparticles with the other catalysts; it is found that the catalysts in this paper have b catalytic performance and can use fewer catalysts in a short time to achieve higher lytic efficiency.

Conclusions
A facile route was used to synthesize non-stacked γ-Fe 2 O 3 /C@TiO 2 double-layer hollow nanoparticles with appropriate overlap mode between the two layers. This unique material structure reduces the catalyst's energy band, which can broaden the light response range to the visible light absorption range, and reduces the recombination rate of photogenerated carriers owing to the C moiety facilitated electron transfer from the γ-Fe 2 O 3 moiety to TiO 2 . The non-stacked γ-Fe 2 O 3 /C@TiO 2 displayed an excellent photocatalytic performance for phenol degradation under visible light irradiation, which is attributed to the cooperativity of the existence of ·OH, ·O 2-, holes. In situ DRIFTS further explored the degradation pathway of phenol; during phenol degradation, -OH was first oxidized, and combined with the identification results of intermediate products, then the benzene rings were destroyed to form small molecule organic acid and eventually mineralized into CO 2 and H 2 O. This overlap mode enhanced both charge generation and charge transfer over photocatalysts. The γ-Fe 2 O 3 moiety endows the nano-shell with an ability of charge generation, while the C moiety facilitated electron transfer from the γ-Fe 2 O 3 moiety to TiO 2 . The unique non-stacked double-layer structure inhibited the initial charge recombination in TiO 2 . The non-stacked γ-Fe 2 O 3 /C@TiO 2 double-layer hollow nanoparticles can make the two steps of charge generation and charge transfer well-matched and synergistically enhanced, which significantly improved the efficiency of the photodegradation of phenol. The preparation of non-stacked γ-Fe 2 O 3 /C@TiO 2 catalysts also provides a novel approach for the future design of a higher-efficiency photocatalytic system.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author upon reasonable request.