Construction of α-Fe2O3/Sulfur-Doped Polyimide Direct Z-Scheme Photocatalyst with Enhanced Solar Light Photocatalytic Activity.

A novel two-dimensional α-Fe2O3/sulfur-doped polyimide (FO/SPI) direct Z-scheme photocatalyst was successfully constructed by a facile thermal treatment method. The effects of α-Fe2O3 nanosheets on the morphology, chemical structure, and photoelectronic properties of FO/SPI composites were systematically characterized by different spectroscopic means. These methods include X-ray diffraction, scanning electron microscopy, X-ray photoelectron spectroscopy, transient fluorescence spectra, and so forth. It was confirmed that the small amounts of α-Fe2O3 can availably facilitate exfoliation of bulk SPI, resulting in a transformation of SPI from bulk to 2D layered composite that illustrates tight interface through the coordination Fe-N bond and an all-solid-state direct Z-scheme junction. Thus, the transfer and separation efficiency of photogenerated electron/hole pairs were significantly enhanced, which greatly promoted improvement of the photocatalytic activity of the FO/SPI composite for methyl orange degradation under solar light. This work provides a new approach to constructing efficient inorganic-organic Z-scheme photocatalyst based on strong interface interaction.

Introduction

The photocatalytic degradation of organic pollutants is the focus of environmental science. It can utilize renewable solar energy to clean serious environmental pollution. The key step of the photocatalytic technology is to develop high efficiency photocatalytic materials. The conjugation polymer-based semiconductors have drawn considerable attention for their low cost, abundant sources, unique chemical stability, and good photocatalytic activity.[1−4] As an outstanding representative of conjugated polymer-based photocatalysts, the graphitic carbon nitride (g-C3N4) has been used in photocatalytic degradation of organic pollutants, water splitting, and reduction of CO2 under visible light irradiation.[5−8] Polyimide (PI) as a special polymer semiconductor, represents another kind of the conjugated polymer photocatalyst that can show excellent visible light response for its chemical stability and suitable band structure.[9] However, the photocatalytic activity of PI on degradation of organic pollutants is still low for the inherent feature of the photonic excitation and carrier transport of conjugated polymer.[10] Therefore, a series of strategies including heterojunctions,[11−14] cocatalysts loading,[15−17] and heteroatoms doping[11,18] have been used to improve the photocatalytic efficiency of PI.

In general, the photocatalytic activity can be improved by reducing the dimensions of the photocatalyst to short diffusion length of photogenerated carriers.[19−22] Moreover, constructing Z-scheme photocatalytic materials can efficiently increase solar-light absorption, separation efficiency of the photogenerated electron–hole pairs, and the redox capacity of the photocatalyst, resulting in a significant improvement of the photocatalytic efficiency.[23−27] Despite the enhanced visible light absorption of sulfur-doped polyimide (SPI) after sulfur doping,[18] its photocatalytic activity remains low due to the fast recombination of photogenerated electron–hole pairs. To promote charge separation and transfer, it is particularly important to construct full-solid-state direct Z-mode composite photocatalytic materials combined with another semiconductor material. Similar to SPI, α-Fe2O3 is also a potential visible-light-driven photocatalyst due to its intrinsic low cost, nontoxic, thermal stability, environmentally friendly, and corrosion-resistant features.[28] Moreover, α-Fe2O3 is an n-type semiconductor with a narrow band gap (about 2.1 eV) and deep valence band (VB). Meanwhile, SPI is also an n-type semiconductor with the higher conduction band bottom level in comparison with α-Fe2O3. Therefore, construction of α-Fe2O3/SPI composite can effectively enhance its charge separation as well as maintain strong redox activity. On the basis of these considerations, we for the first time constructed a novel 2D α-Fe2O3/sulfur-doped polyimide (FO/SPI) Z-scheme photocatalyst for the degradation of methyl orange (MO) through solar-light driven. During the preparation of FO/SPI composites, α-Fe2O3 as a catalyst promotes the oxidation stripping of bulk SPI to form 2D layered FO/SPI composites similar to g-C3N4, which were confirmed by XRD and SEM. The PL spectral of FO/SPI samples illustrated that the small amount of α-Fe2O3 nanosheets can obviously suppress the recombination of photogenerated electron–hole pairs because of their Z-scheme band structure. The effects of α-Fe2O3 nanosheets on the solar-light photocatalytic degradation efficiency of FO/SPI composites were systematically investigated. The possible photocatalytic mechanism was also proposed.

Experimental Section

Catalyst Preparation

Melamine (MA) and pyromellitic dianhydride (PMDA) were bought from Shanghai Macklin Biochemical Co., Ltd., whereas iron trichloride (FeCl3), sublimed sulfur (S4), propanetriol (C3H8O3), and urea (CH4N2O) were purchased from Tianjin Dengke Chemical Reagent Co., Ltd. Triethanolamine (TEOA), KI, C4H10O (TBA), as well as thiamethoxam (TM) were procured from Sigma-Aldrich. All chemical reagents were used without further purification.

Sulfur-doped polyimide (SPI) was synthesized through a solid–phase thermal polymerization method as previously reported.[18] In a typical procedure, the mixture of 1.26 g of MA, 2.18 g of PMDA, and 25 w.t% of sublimed sulfur (S4) was put into a semiclosed porcelain crucible and heated to 325 °C with a heating rate of 7 °C min–1 using through-nitrogen gas. After keeping this temperature for 4 h under N2 flow, the resulting solid was cooled to room temperature and milled into powders.

α-Fe2O3 was synthesized through a modified method based on previous literature.[29,30] Typically, the mixture of 4.87 g of FeCl3·6H20 and 4.32 g of CH4N2O was dispersed in glycerol (10 mL) and deionized water (60 mL) under magnetic stirring. Next, the stirred solution was transferred to a Teflon-lined stainless steel autoclave with an 80 mL capacity and maintained at 180 °C for 12 h. After cooling to room temperature, the obtained mixture was separated by centrifugation and washed with deionized water several times. Then, it was dried at 80 °C for 12 h and further heated to 550 °C for 2 h in an air atmosphere. The obtained gray-black powdered solid was α-Fe2O3 after grinding.

The α-Fe2O3/sulfur-doped polyimide composites were fabricated by a method similar to that of SPI. The preparation of 5%α-Fe2O3/SPI was taken as an example. The mixture of 0.95 g of SPI and 0.05 g of α-Fe2O3 was put in a porcelain boat and heated to 325 °C for 4 h before cooling to room temperature. The obtained solid was milled and named as 5%α-Fe2O3/SPI and abbreviated as 5FO/SPI. Therefore, the composites of 10%α-Fe2O3/SPI, 12%α-Fe2O3/SPI, 17%α-Fe2O3/SPI, 20%α-Fe2O3/SPI, and 28%α-Fe2O3/SPI with different contents of α-Fe2O3 were prepared and abbreviated as 10FO/SPI, 12FO/SPI, 17FO/SPI, 20FO/SPI, and 28FO/SPI, respectively. The amounts of the reactants for synthesis of the FO/SPI composite samples are summarized in Table S1 of the Supporting Information.

Characterization

Powder X-ray diffraction (PXRD) measurements of the samples were performed on a Rigaku diffractometer using Cu Kα radiations, and the X-ray tube was operated at 40 kV and 40 mA. FTIR measurements were obtained by using a Nicolet 6700 spectrometer using the KBr pellet support. The scanning electron microscope (SEM) image was recorded on a JSM-6610 system. The transmission electron microscope (TEM), high-resolution transmission electron microscope (HRTEM) was collected with a JEM-2100 electron microscope. Brunner-Emmet-Teller’s (BET) method and Barrett–Joyner–Halenda’s (BJH) method was taken into consideration to inspect the curves of the surface area and pore size distribution, respectively. UV–vis diffuse reflection spectroscopy (DRS) was taken on a Shimadzu UV–2600 spectrometer using BaSO4 as the reference at room temperature. The photoluminescence (PL) spectrum was performed by an Agilent MY15170004 spectrometer with an excitation wavelength of 350 nm. X-ray photoelectron spectroscopy (XPS) and valence band X-ray photoelectron spectroscopy (VBXPS) were conducted on a PHI 5000 Versa Probe X-ray photoelectron spectrometer with monochromatized Al Ka X-ray radiation. Electron paramagnetic resonance (EPR) measurement was tested by a Bruker EMX-10/12 EPR spectrometer at room temperature. The transient fluorescence spectra of the photocatalysts were recorded by an FLS980 multifunction steady-state and transient fluorescence spectrometer (Edinburgh Instruments) at room temperature.

Electrochemical Measurements

The photoelectrochemical properties are measured on the traditional three-electrode system. It was made up of the working electrode (the glass with sample), the counter electrode (a Pt sheet), a reference electrode (an Ag/AgCl electrode), and the electrolyte solution (Na2SO4, 0.5 mol L–1, pH = 6.8). The working electrode was prepared on fluorine-doped tin oxide (FTO) transparent conductive film glass by electrophoretic deposition method. Typically, the mixture of 50 mg of powder sample and 12 mg of I2 was dispersed in 45 mL of acetone under sonication for 25 min to obtain the turbid liquid. Meanwhile, two pieces of the FTO transparent conductive film glass that have been arranged on the electrophoretic deposition apparatus were immersed in the turbid liquid, turning on the switch and keeping the voltage at 21 V for 6 min. One piece of the glass with the certain of sample on the conductive face was obtained and dried naturally in air. In order to reduce the influence of the thickness of sample layer and the size of the light irradiation area, the light irradiated on the back side (FTO substrate/semiconductor interface) and the exposure area of each working electrode was 0.28 cm2. The electrochemical experiments were performed on an electrochemical analyzer (CHI-663C, Shanghai Chenhua, China). Mott–Schottky curves were taken under dark conditions at 1.0 kHz frequencies. For Nyquist plots measurements, the frequency ranged from 200 kHz to 10 mHz, and the perturbation signal was10 mV.

Photocatalytic Performance Tests

The photocatalytic activity of the samples was evaluated by the degradation of MO under full arc light irradiation. MO (99.7% purity) was purchased from Tianjin Kemeio Chemical Reagent Co., Ltd. The experimental device for photocatalytic degradation consists of photocatalytic glass reactor, xenon lamp, cryotemperature tank, and magnetic mixer. Light source was achieved by a 300 W xenon lamp (I = 20 A) and a cooling fan. A photocatalytic reactor was a Pyrex top-irradiation full glass container with a constant temperature water system for holding temperature at 25 °C. In each test, 0.2 g of catalyst was dispersed in 100 mL aqueous solution of MO (4 mg L–1), and the suspension was magnetically stirred in the dark for 1 h to reach the adsorption–desorption equilibrium before light irradiation. At predetermined time intervals, 4 mL of suspension was collected and centrifuged to remove most of the catalyst particles. Then the obtained clear solution was filtered through MCE 0.45 μm and analyzed by measuring its absorbance with an UV–vis spectrometer (Mepoda UV-1800) at 464 nm. Using the same method, the photocatalytic degradation activitives of the TM as a colorless pollutant besides the dye were tested. The only difference is that the measured absorption wavelength was at 310 nm.

Recyclability of FO/SPI Composites

To further evaluate the cyclic stability of photocatalytic materials, experiments were conducted for 6 runs, and the irradiation time of each run was 180 min. The used photocatalyst could be conveniently filtered and collected from the solution by centrifugation and added back into the reactor for the next run.

Results and Discussion

Structure and Morphology Analysis

The XRD patterns of pristine SPI, α-Fe2O3, and FO/SPI composites with different α-Fe2O3 contents were illustrated in Figure . Several distinct peaks in the range of 10–30° appeared on the pattern of pristine SPI, which is consistent with the XRD diffraction peaks of SPI reported in the previous literature.[18] Meanwhile, the XRD characteristic peaks of α-Fe2O3 (JCPDF: 52–1449) can be obviously observed in the FO/SPI composites. The intensities of two peaks (102) at 33.1° and (110) at 37.2° and other peaks attributed to α-Fe2O3 gradually increasing for increased loadings of α-Fe2O3 in the composites, confirming that the α-Fe2O3 was indeed introduced into the SPI. Amazingly, most of the XRD characteristic peaks of SPI disappeared after combining with α-Fe2O3 nanosheets. This may be because the SPI was oxidized and stripped to form a 2D-layered material by α-Fe2O3 as a catalyst[31] during the synthesis of the FO/SPI composite. The weak characteristic peak located at ∼13° in the XRD spectrum of the conjugated polymer is attributed to an in-plane repeat period of 0.861 nm in the crystal.[1] It has vanished after introducing α-Fe2O3 into the SPI to form FO/SPI hybrids, presumably because of the strong interaction between α-Fe2O3 and SPI, thus affecting the formation of an in-plane repeat period in the preparation of FO/SPI hybrid photocatalysts. Moreover, the strong distinct peak centered about ∼18° and ∼29.7°, respectively, representing the polymerization degree of the SPI[16] have disappeared, but the characteristic peak at ∼27.4° reflecting the stacking of π-conjugated two-dimensional (2D) frameworks of SPI still exist in the XRD patterns of FO/SPI composites. These may be attributed to the special thermal preparation process of FO/SPI composites. In detail, the SPI first was oxidized and stripped under the catalysis of α-Fe2O3 to form some oligomeric polyimides,[31] and then these oligomers further repolymerized with the aid of the π–π electronic interaction or hydrogen bonding between the conjugated core units, resulting in an assembly of highly regular and symmetrical polymer frameworks.[18] Thus, the more orderly layered 2D-FO/SPI hybrids formed through a new preparation method of 2D FO/SPI composites including exfoliation and further repolymerization.

Figure 1

XRD patterns of pristine SPI, α-Fe2O3, and FO/SPI composites with different α-Fe2O3 contents.

To further determine the morphology of the prepared samples and the existence form of α-Fe2O3 on SPI, SEM, TEM, and EDS measurements were carried out. As shown in Figure a, most of α-Fe2O3 nanosheets crystals displayed uniform hexagonal morphology with the thickness of ∼50 nm. HR-TEM images obviously demonstrate a lattice spacing of ∼0.25 nm, which matches well with the (110) plane of α-Fe2O3 crystals (Figure c). The SPI sample presents porous and rough bulk morphology (Figure S1). But the layered morphology of the 17FO/SPI composite sample can be clearly observed, indicating the transformation of SPI from bulk to 2D layers of FO/SPI composite as shown in Figure b. Meanwhile, the junctions of α-Fe2O3 nanosheets crystals/2D layers SPI composite were successfully constructed, which was evidenced by the observation of sharp interfaces between α-Fe2O3 nanosheets and 2D layers SPI in 17FO/SPI sample with α-Fe2O3 loading (Figure d). The energy dispersive X-ray (EDX) elemental mapping of the 17FO/SPI composites on a microscope further suggests uniform distribution of α-Fe2O3 over 2D SPI surface (Figure e and Figure S2). In addition, the BET-specific surface area and BJH pore size distributions of the as-prepared samples were assessed by the N2 adsorption–desorption test. Obviously, each sample exhibited a type IV with a H3 hysteresis loop, illustrating the presence of a mesoporous structure within the samples (Figure S3). The pore size distributions of all samples were predicted by the BJH method (Figure S3b of the Supporting Information). The average pore diameter, specific surface area, and pore volume of all samples were measured by the BET method (Table S2). Pure α-Fe2O3 has a low surface area but the SPI and FO/SPI composites have appropriate surface area for efficient absorption as well as degradation. Thus, the SPI acts as the support for α-Fe2O3 to enhance its surface area.

Figure 2

Images of SEM (a,b) and TEM (c,d) for α-Fe2O3 and 17FO/SPI composite, and EDX elemental mappings of C, N, S, O, and Fe for the 17FO/SPI composite (e).

The interface interaction of FO/SPI is verified by FTIR spectroscopy (Figure ). As shown in the FTIR spectrum, all of the FO/SPI composite samples exhibit the characteristic absorption bands of sulfur-doped polyimide.[18] The bands at 1787.5, 1723.5, and 726.3 cm–1 are ascribed to the asymmetric stretching, symmetric stretching, and bending vibrations of the −C=O bond in the PMDA moiety of SPI, respectively. The bands at 1553 and 1308 cm–1 marked with blue dotted lines are indicative of the stretching vibration of C–N–C in the five-membered imide rings and breathing modes of aromatic C=N in the triazine unit, respectively.[32] The weak characteristic vibrating peak of the S–N bond at 632 cm–1 marked with a red dotted line[33] was also observed in a series of samples. These results indicate that the main bonding skeleton or chemical framework of SPI has not been changed after introducing the Fe species into SPI. However, the broad peaks at around 3100–3400 cm–1 assigned to the typical amine groups, hydrogen bonds, and water molecule were observed that it became weak after loading α-Fe2O3 on SPI.[33] It may have attributed to the partial stripping of layered structure of SPI during the solid phase thermal synthesis of FO/SPI composites, because of the multiple O–H bond interactions of interlayer and absorbed H2O were lost at 352 °C, which matches well with XRD results. Meanwhile, the infrared characteristic absorption peaks of α-Fe2O3 were not observed in a series of FO/SPI samples and may be attributed to the very weak absorption strength of iron oxide and its high dispersion on the SPI surface. In addition, it is surprising that a new peak at 1403 cm–1 appeared in the spectra of all FO/SPI composites, which is assigned to the characteristic peaks of the Fe–N bond.[34] The formation of the Fe–N bond is likely to arise from coordination of the N atoms of heptazine units to the unsaturated Fe sites of α-Fe2O3 in the interface of FO/SPI. This Fe–N bond confirmed that there is a strong chemical interaction between α-Fe2O3 and SPI, thus leading to construct an intimate contact interface. The intimate contact interface between α-Fe2O3 and SPI can promote the transfer of the photogenerated carriers.

Figure 3

FT-IR spectra of α-Fe2O3, SPI, and FO/SPI composites.

To further pursue the surface chemical state of the elements in the involved samples, XPS spectra of the as-prepared SPI, α-Fe2O3 and 17FO/SPI composite were recorded at room temperature. The binding energies of the elements were calibrated by the C 1s peak (284.6 eV), arising from adventitious carbon. The corresponding high resolution XPS spectra of C 1s, N 1s,

Fe 2p and O 1s of the photocatalysts are shown in Figure , respectively. The C 1s XPS spectra of SPI are obviously observed from Figure a. In three C 1s peaks located at 284.6, 286.6, and 288.7 eV, the C 1s peak at 284.6 eV was ascribed to the sp2 C=C bonds or adventitious carbon, while the peaks centered at 286.6 and 288.7 eV could be attributed to the N–C–N bond and the C=O in the triazine rings of SPI, respectively. For the N 1s XPS spectra of SPI and 17FO/SPI composite, it could be shown in Figure b. The N 1s peak at ∼397.8 eV corresponds to the sp2–bonded N in the triazine rings (N–C=N) of SPI. At the same time, a pairs of peaks at the binding energies of 398.8 and 399.2 eV are stemmed from the splitting of the peak of nitrogen atoms in the five-membered imide ring of polyimide (PI) after incorporating sulfur into the triazine ring of PI. These results are matched well with the previous literature values.[17] By more carefully comparing the spectrum of 17FO/SPI composite with that of the pristine SPI, it can be clearly observed that the binding energies of N 1s for the 17FO/SPI at 397.9, 399.1, and 399.5 eV shifted toward high binding energies. In addition, the binding energies of Fe 2p for α-Fe2O3 are determined to be 710.6 and 723.7 eV as shown in Figure c, which matched well with reported values of the Fe 2p3/2 and Fe 2p1/2, ascribing to 3+ oxidation states of iron.[35] Obviously, the binding energies of Fe 2p for the 17FO/SPI composite at 710.3 and 723.1 eV corresponding to the Fe 2p3/2 and Fe 2p1/2 shifted toward lower binding energies compared to that of pure α-Fe2O3. As previously reported, the sp3-hybridized N atoms in tertiary nitrogen groups of g-C3N4 can provide the lone pair electrons to the unoccupied d orbital of metal atoms to form M (δ +)–N (δ –) bonding states through chemical coupling interaction.[36] Intriguingly, the similar interaction also exits in our FO/SPI composite. The lone pair electrons of N atoms in the integral conjugated framework of SPI interact with the unoccupied d orbital of Fe atoms in α-Fe2O3, leading to the increase of N 1s binding energy and the reduction of Fe 2p binding energy in 17FO/SPI composite compared that of pristine SPI and α-Fe2O3. Thus, the interaction between α-Fe2O3 nanosheets and SPI is consistent with the formation of Fe–N bond in the infrared spectrum of the 17FO/SPI composite sample. However, the binding energy of O 1s in the high-resolution O 1s XPS spectra did not obviously changed as shown in Figure d. The peak at 530.3 eV is attributed to lattice oxygen in α-Fe2O3, and the peak at 532.5 eV is assigned to surface −OH groups/adsorbed water.[37] Besides these two peaks, a new O 1s peak centered at 531.4 eV was observed in the XPS spectrum of the 17 FO/SPI composite, which is assigned to the C=O in the five-membered imide ring of SPI.[38] For the unchanged binding energies of O 1s after introducing α-Fe2O3 on SPI, this may be ascribed to the higher electronegativity of an oxygen atom than that of a nitrogen atom. Thus, the stronger chemical interaction between SPI and α-Fe2O3 was promoted to form covalent Fe–N bonding states. The formation of Fe–N bonding states benefits the transport and separation of photogenerated carriers between polymer supports and transition metal oxide leading to a great improvement in photocatalytic activity and stability of inorganic–organic composite semiconductor materials. In addition, the electron paramagnetic resonance (EPR) was conducted to further reveal the interaction between α-Fe2O3 and SPI at room temperature. From Figure , it can be observed that the Lorentzian line centered at a g value (2.003) is detected on the spectra of SPI and 17 FO/SPI composite powder samples in the magnetic field from 3000 to 4000 G. The peak signal at g value equivalent to 2.003 is attributed to an unpaired electron on the carbon atoms of the aromatic rings within π-bonded nanosized clusters.[39] Moreover, the EPR intensity of the peak at g = 2.003 of the 17 FO/SPI composite clearly strengthened more than that of SPI, implying that the extension of the conjugate system sharply enhanced after introducing α-Fe2O3 into SPI.

Figure 4

XPS spectra for (a) C 1s, (b) N 1s, (c) Fe 2p, and (d) O 1s of SPI, α-Fe2O3, and 17FO/SPI composite.

Figure 5

EPR spectra of SPI, α-Fe2O3, and 17FO/SPI composite samples at room temperature.

Optical and Electronic Properties

The optical absorption properties of SPI, FO/SPI, and α-Fe2O3 samples were investigated by UV–vis diffuse reflection spectra and converted to Kubelka–Munk function. As shown in Figure a, after combining with α-Fe2O3 nanosheets the absorption band edge of SPI is obviously red-shifted and the light absorption intensity of FO/SPI samples significantly enhanced with the increasing concentration of α-Fe2O3. The bulk SPI possessed an absorption edge at 475 nm and a corresponding band gap of 2.53 eV (Figure b), while the absorption edge of 17FO/SPI sample red-shifted to 510 nm, resulting in a band gap of 2.35 eV (Figure S4). This optical property of UV–vis spectra shows the generation of new localized electronic states (surface states) within the band gap of SPI,[40] which may induce the absorption of visible light. Therefore, the absorption intensities of FO/SPI samples were remarkably enhanced in the range of 400–800 nm. The increased visible light absorption and the decreased band gap of the composites are beneficial to the production of photogenerated electrons and holes, thus promoting the enhancement of their photocatalytic activity. Moreover, the bandgap of the prepared samples can be calculated by the Kubelka–Munk function equation, αhv = A(hv – Eg)1/2.[34] From Figure b, it can be clearly observed that the band gaps of SPI and α-Fe2O3 samples were approximately 2.53 and 2.05 eV, respectively, which obtains from extrapolation of the linear region of the absorbance squared versus energy. Meanwhile, the VBXPS spectra of SPI and α-Fe2O3 samples displayed that the energy levels of valence band (VB) in SPI and α-Fe2O3 are about 1.64 and 2.31 eV, respectively, as shown in Figure c. In addition, combining these findings with the UV–vis and VBXPS results, the conduction band (ECB) potentials are −0.89 and 0.26 eV for SPI and α-Fe2O3 versus NHE calculated by the equation ECB = EVB – Eg.[41] The conduction band (ECB) potentials can also be estimated by the Motshoky curves. As shown in Figure S5, the flat band potentials (Efb) of SPI and 17FO/SPI are −0.89 and −1.08 V versus Ag/AgCl, respectively. It is well-known that the CB level is generally more negative about −0.1 or −0.2 V than its flat band potential for n-type semiconductor.[34] Thus, the ECB for SPI and 17FO/SPI is presumably −0.89 and −1.08 V, respectively. Therefore, the CB level of SPI changed significantly after binding to α-Fe2O3. So the schematic band structure of SPI and α-Fe2O3 was determined using these results as illustrated in Figure d.

Figure 6

(a) UV–vis DRS spectra of as-prepared pure SPI, α-Fe2O3, and 17FO/SPI composites. (b) The corresponding plots of (αh ν)2 versus h ν, (c) VBXPS spectra, and (d) schematic illustration of band structures for SPI and α-Fe2O3.

To reveal the efficiency of charge carrier trapping, transfer, and separation in prepared samples, photoluminescence (PL) measurement was carried out, as shown in Figure a. After combining with α-Fe2O3 nanosheets, the fluorescence intensities of the FO/SPI composite powder samples sharply decreased compared with that of the SPI powder sample. This result indicates that the construction of Z-scheme FO/SPI efficiently suppressed the radiative recombination of the photogenerated electrons and holes, which is mainly attributed to the excellent charge carriers transport principle of Z-scheme FO/SPI interface structure between α-Fe2O3 nanosheets and 2D SPI and reduced dimension of FO/SPI semiconductors. In addition, the time-resolved photoluminescence (TRPL) decay spectra of SPI, 17FO/SPI, and α-Fe2O3 samples were exhibited in Figure b. The lifetimes about τ1 and τ2 of the 17FO/SPI composite are between SPI and α-Fe2O3, respectively. It may be caused by the combination of SPI and α-Fe2O3. The average lifetime τav are 3.47, 3.17and 9.08 ns for SPI, 17FO/SPI and α-Fe2O3, respectively. The result showed that the lifetime of excited state of the 17FO/SPI composite was shorter than that of SPI and α-Fe2O3.[42] Moreover, the decreased lifetime of photoexcited charges was concerned with promoted electron transport.[34]

Figure 7

(a) Comparison of photoluminescence (PL) spectra of SPI, FO/SPI composites, and α-Fe2O3 samples. (b) Transient PL decay spectra for pure SPI, 17FO/SPI and α-Fe2O3..

To further investigate the electrochemical conductivity properties of the prepared samples, EIS as a powerful tool was used. The impedance of the EIS Nyquist plot is the impedance of the copper foil/photocatalyst, photocatalyst/photocatalyst, and photocatalyst/electrolyte interfaces. The lower the impedance is, the lower the interface resistance is and the higher the separation efficiency of the photogenerated carriers is.[43−45] As displayed in Figure a, the semidiameter of the Nyquist curve for 17FO/SPI composite was much smaller than that of SPI, showing the FO/SPI composite has better electron transport capability than SPI alone. Meanwhile, the photocurrent response was then performed to further study the charge transfer behavior of the prepared samples.[46] The photocurrent responses of 17FO/SPI and SPI after deposition on FTO electrodes versus Ag/AgCl under full arc light irradiation are shown in Figure b. It can be clearly observed that the photocurrent density over 17FO/SPI is significantly enhanced, which is about 5.0 times as high as that of bulk SPI. Ordinarily, the formation of the photocurrent is mainly attributed to the diffusion of the photogenerated electrons to the back contact and simultaneously the photoinduced holes are taken up by the hole acceptor in the electrolyte. Therefore, the enhanced photocurrent of 17FO/SPI sample indicates a longer lifetime of the photoinduced charge carriers and more efficient separation of photogenerated electron/hole pairs than that of bulk SPI sample, which is consistent with PL analysis and beneficial for its improved photocatalytic activity.

Figure 8

(a) EIS Nyquist plots of electrochemical impedance spectroscopy with SPI and 17FO/SPI composite in the dark. (b) Photocurrent–potential curves of SPI and 17FO/SPI electrode in the 0.5 mol L–1 Na2SO4 aqueous solution under full arc light irradiation.

Photocatalytic Activity and Mechanism

The photocatalytic activities of SPI, α-Fe2O3 and FO/SPI composite powder samples are examined by the degradation of MO in solution (4 mg L–1). Usually, the initial concentrations of MO are identical for all the samples, while the concentration of MO at the time as the light was turned on did differ (in Figure a) because the specific surface area varied for each sample (Table S2). As shown in Figure a, the photolysis of MO under full arc light (simulation of solar light) irradiation is negligible in the absence of the photocatalyst. Meanwhile, little degradation of methyl orange was observed on the 17FO/SPI composite sample in dark condition even for 5 h due to the stable structure of MO. However, the prepared samples displayed different degradation activity for MO under the irradiation of a xenon 300 W lamp without a cut-off filter, demonstrating that the degradation of MO was indeed driven by full arc light. The direct correlation between the activity of the 17FO/SPI composite sample and the energy of irradiated light was further confirmed by the wavelength dependence experiment using different pass-filters. From Figure b, it can be observed that the degradation activity trend is consistent with its absorption edge, which illustrates that the reaction of photocatalytic degradation is really driven by the bandgap transition of the 17FO/SPI composite sample. Obviously, the photocatalytic activities of all FO/SPI composites are higher than those of pristine SPI and α-Fe2O3, indicating the close binding of α-Fe2O3 and SPI not only extends absorption for visible light but also promotes the separation of photogenerated electrons and holes, thus leading to a significant improvement of the photocatalytic activity on SPI sample. Moreover, with the increase of α-Fe2O3 content on SPI the photocatalytic degradation activity of FO/SPI composites increases and reached a maximum for the 17FO/SPI composite sample, which is about 3.8 times higher than that of SPI in 2 h. Then, the degradation activity of 20FO/SPI and 28FO/SPI composite samples gradually decreased by increasing the α-Fe2O3 content on SPI. As for the drop in the activity of the samples with heavy loadings of α-Fe2O3, it is likely due to the insignificant increase of photogenerated electrons and holes separation efficiency compared to that of the 17FO/SPI composite sample. As shown in Figure , the PL spectra lines of 17-, 20-, and 28FO/SPI composite samples almost overlap, indicating adding more α-Fe2O3 in SPI does not cause an increase in the separation efficiency of photoelectron/hole pairs. In addition, the effect of pH on the photocatalytic activity of catalyst is indeed a key factor in determining the photocatalytic reaction.[47] We performed a series of experiments to determine the photocatalytic degraded MO activities of the 17FO/SPI composite sample at the distinct pH values like 1, 3, 5, 7, 9, and 11 acquired by the subsequent addition of NH3·H2O or HCl. As shown in Figure S6a, The photocatalytic degradation methyl orange activity of the 17FO/SPI composite first enhances then decreases with increasing pH values. Obviously, the highest degrading efficiency of 17FO/SPI was determined to be 87.9%, with rate constants of 0.7 h–1 at PH = 7 (Figure S6b). To compare the photocatalytic activity of FO/SPI composite with other systems, we prepared g-C3N4 according to the literature method,[6] and then prepared the 17 wt %α-Fe2O3/g-C3N4 (17FO/CN) sample by using a synthetic method of 17FO/SPI sample. It is clearly observed that the photocatalytic degradation MO activity of 17FO/SPI is better than 17FO/CN under the same photocatalytic reaction conditions (Figure S7). In order to avoid color sensitivity, the photocatalytic degradation activities of the TM as a colorless pollutant besides dye were tested. Apparently, the photolysis of TM under full arc light irradiation is negligible in the absence of photocatalyst. However, the prepared the 17FO/SPI sample displayed the good degradation activity for TM under the light irradiation (Figure S8). It is shown that the degradation of TM is indeed a photocatalysis of FO/SPI composite sample but not color sensitivity. To further study the effects of the amount of photocatalyst and the initial concentration of pollutants on the photocatalytic activities, the contrast experiments were carried out. As shown in Figure S9a, the photocatalytic activity of the 17FO/SPI composite increases with the increasing amount of the catalyst. However, the excessive amount of the catalyst did not significantly enhance its photocatalytic activity. In addition, the change in the initial pollutant concentration did not cause significant differences in the photocatalytic activity of 17FO/SPI composite (Figure S9b).

Figure 9

(a) Comparison of the photocatalytic degradation of MO over different catalysts under full arc light (λ > 300 nm). (b) Dependence of degradation activity on wavelength by 17FO/SPI.

For a good photocatalyst, another important criterion is high durability. To investigate the cyclic stability of a prepared catalyst, photocatalytic degradation of MO was continuously conducted on the 17FO/SPI composite sample for 1080 min in six cycles. As displayed in Figure a, it can be clearly observed that there is no obvious abate in the photocatalytic degradation activity of the 17FO/SPI sample after 6 runs. Meanwhile, its powder XRD pattern is similar before and after photocatalytic reaction of MO degradation (Figure b). These results demonstrate that the FO/SPI composite sample has satisfactory recyclability and stability for the photocatalytic degradation of organic pollutants.

Figure 10

(a) Cycling runs for the photodegradation of MO in the presence of the 17FO/SPI composite sample under full arc light illumination. (b) XRD patterns of the 17FO/SPI before and after photocatalytic degradation of MO under full arc light illumination.

To further inquire about the photocatalytic degradation MO mechanism of the FO/SPI composite in detail, a series of controlled experiments were conducted in the different reaction conditions. Normally, photogenerated electrons can reduce oxygen to form reactive superoxide radical anion (·O2–), and photogenerated holes can react with water to produce hydroxyl radicals (·OH).[48] Therefore, through incoming nitrogen gas to removal of oxygen experiments were first performed to determine the effect of photogenerated electrons on photocatalytic degradation MO activity of FO/SPI composite. From Figure a, it is clearly observed that the photocatalytic activity of 17FO/SPI significantly reduced under N2, illustrating photogenerated electrons that play a key role. Moreover, since the photogenerated holes can oxidize I– into I2, the experiment using KI as a hole sacrificial agent was conducted to determine the effect of holes on photocatalytic activity. Minor decrease of photocatalytic activity on the 17FO/SPI composite sample was observed after the addition of KI (Figure a), indicating the photogenerated holes do not play a leading role in photocatalysis. This result was further confirmed by using C4H10O (TBA) and triethanolamine (TEOA) as ·OH and h+ sacrificial agents, respectively (Figure a). As is well-known, the ·OH/H2O potential is 2.27 eV,[49] and the ·O2–/O2 potential is −0.28 eV.[50] From Figure , we can see that the valence band position (1.64 eV) (Figure d) of SPI is not positive enough to oxidize H2O into ·OH (2.27 eV),[51] but it can reduce the absorbed O2 to form ·O2– because that the CB level of SPI at −0.89 eV (Figure d) is more negative than the ·O2–/O2 potential (−0.28 eV). Meanwhile, it is also observed that the VB and CB levels of α-Fe2O3 are 2.31 and 0.26 eV (Figure d), respectively. So the photogenerated holes (h+) on the valence band of α-Fe2O3 can oxidize H2O to yield ·OH, and its photogenerated electrons (e–) cannot react with O2 to form ·O2– because its CB level (0.26 eV) is not negative enough to drive the reduce process of O2 to form ·O2–.[52] These analyses are further confirmed by the ESR spin-trapping using 5,5-dimethyl1-pyrroline N-oxide (DMPO) as a spin trap to stabilize radicals. To detect the transient radical intermediates, ESR was performed in aqueous solution (for DMPO-·OH) and methanol solution (for DMPO-·O2–),[53] respectively. From Figure b, it can be observed that no signals of DMPO-·OH and DMPO-·O2– were detected from α-Fe2O3, SPI, and 17FO/SPI hybrid samples without light irradiation.[54] As shown in Figure c, the strong signals of DMPO-·OH and DMPO-·O2– were detected from α-Fe2O3 and SPI under light irradiation, respectively. Meanwhile, no obvious signals of ·OH and ·O2– were observed in the SPI and α-Fe2O3 samples under the same condition, respectively. Significantly, the obvious ·OH and ·O2– signals were simultaneously detected in the presence of the 17FO/SPI composite (Figure d). Moreover, the strong signals of ·OH and ·O2– were also simultaneously detected in the presence of the 5FO/SPI and 28FO/SPI composites under full arc light irradiation (Figure S10). Thus, if it is a traditional type-II heterojunction[55] between α-Fe2O3 and SPI (Figure ), the holes of α-Fe2O3 should transfer to the VB of SPI, which cannot oxidize H2O into ·OH and will lead to a more serious photocorrosion. Usually, photocorrosion happens due to the sulfide being oxidized by holes. If so, it is unlikely that the FO/SPI composite can have a good photocatalytic stability. At the same time, the photogenerated electrons (e–) of SPI transfer to the CB of α-Fe2O3, which cannot react with O2 to form ·O2–. Therefore, it should be a direct Z-scheme route[56,57] of photogenerated carriers existing in the FO/SPI composite photocatalyst.

Figure 11

(a) The effects of TEOA, KI, TBA, and purging-N2 as h+, ·OH, and ·O2– sacrificial agents on the photocatalytic degradation of MO in the presence of the 17FO/SPI composite under full arc light irradiation. (b) DMPO spin-trapping ESR spectra of α-Fe2O3, SPI, and 17FO/SPI composite without light irradiation. (c,d) DMPO spin-trapping ESR spectra of α-Fe2O3, SPI, and 17FO/SPI composite (in aqueous for DMPO-·OH; in methanol for DMPO-·O2–).

Figure 12

Schematic illustration of the traditional type-II heterojunction and the proposed Z-scheme charge transfer mechanism.

On the basis of the above results, we proposed a Z-scheme photocatalytic mechanism for the enhanced activity and stability of FO/SPI composite as shown in Figure . Under full arc light irradiation, both SPI and α-Fe2O3 were excited to yield photogenerated electron (e–)/hole (h+) pairs. Then the photoexcited e– in the CB of α-Fe2O3 easily recombine with the photoexcited h+ on the VB of SPI, resulting in numerous more active photogenerated e– and h+ remaining in CB of α-Fe2O3 and VB of SPI, respectively. The remaining e– in the CB of SPI can easily be captured by O2 to form ·O2– while the h+ left behind in the VB of α-Fe2O3 can react with H2O to yield ·OH.[58,59] Both active species ·O2– and ·OH can easily oxidize MO to form inorganic small molecules, such as CO2, H2O, and so forth.[60,61] Thus, an effective process of photocatalytic degradation of MO proceeds smoothly.

Figure 13

Schematic of the photogenerated charge carrier’s separation and transfer in the FO/SPI system under full arc light irradiation.

Conclusions

In summary, we adopt a facile thermal treatment method to successfully construct a novel 2D direct Z-scheme photocatalyst. The small amounts of α-Fe2O3 can availably facilitate exfoliation of bulk SPI, resulting in a transformation of SPI from bulk to 2D layered hybrid, which illustrates tight interface and an all-solid-state Z-scheme junction through the coordination Fe–N bond. Thus, the transmission and separation efficiency of photogenerated electron–hole pairs were significantly enhanced, which greatly promoted improvement of the activity and durability of the FO/SPI composite for MO degradation under solar light.