Self-cleaning semiconductor heterojunction substrate: ultrasensitive detection and photocatalytic degradation of organic pollutants for environmental remediation

Emerging technologies in the field of environmental remediation are becoming increasingly significant owing to the increasing demand for eliminating significant amounts of pollution in water, soil, and air. We designed and synthesized MoS2/Fe2O3 heterojunction nanocomposites (NCs) as multifunctional materials that are easily separated and reused. The trace detection performance of the prepared sample was examined using bisphenol A (BPA) as the probe molecule, with limits of detection as low as 10−9 M; this detection limit is the lowest among all reported semiconductor substrates. BPA was subjected to rapid photocatalytic degradation by MoS2/Fe2O3 NCs under ultraviolet irradiation. The highly recyclable MoS2/Fe2O3 NCs exhibited photo-Fenton catalytic activity for BPA and good detection ability when reused as a surface-enhanced Raman scattering (SERS) substrate after catalysis. The SERS and photocatalysis mechanisms were proposed while considering the effects of the Z-scheme charge-transfer paths, three-dimensional flower-like structures, and dipole–dipole coupling. Moreover, the prepared MoS2/Fe2O3 NCs were successfully applied in the detection of BPA in real lake water and milk samples. Herein, we present insights into the development of MoS2/Fe2O3 materials, which can be used as multifunctional materials in chemical sensors and in photocatalytic wastewater treatments for the removal of recalcitrant organic pollutants.


Introduction
Serious environmental pollution and accelerated global warming are attributed to the rapid consumption of fossil fuels, the increasing population, and the rapid development of the economy. Thus, the development of innovative and renewable environmental remediation materials is becoming increasingly important 1-6 . Since mechanically exfoliated graphene was discovered, the development of two-dimensional (2D) materials consisting of atomically thin crystal layers bound by van der Waals forces has accelerated owing to the potential applications of these materials in optoelectronics, catalysis, new technologies, and electricity [7][8][9] . 2D-MoS 2 nanosheets are excellent layered materials, having unique layered structures and large surface areas. It is important to investigate methods for improving the chemical properties of MoS 2 , which may affect its application in electronic devices, catalysis, and molecular sensing 10 . A popular method for improving the properties of MoS 2 is the decoration of MoS 2 with noble metal nanoparticles. For instance, a MoS 2 /noble metal nanoparticle composite can induce local surface plasmon resonance (LSPR) for activating the photoelectrocatalysis of H 2 and enhancing the light absorption or emission of MoS 2 . Moreover, the LSPR can generate surface-enhanced Raman scattering (SERS), which can be used in biological and chemical sensing applications [11][12][13] . Among various traditional noble metal materials, Au nanomaterials are the most widely used SERS substrate material [14][15][16] . However, the high cost and specialized instruments required for Au substrates hinder their practical application. Due to its very high SERS activity, Ag is another widely studied substrate material 17 . Although the price of Ag is much lower than the price of Au, the main defect of Ag is its poor stability, which easily oxidizes in air. To address these problems, it is essential to exploit synergistic effects by incorporating inexpensive and stable semiconductors.
Investigations have been continuously conducted on the efficient separation of a nanocomposite (NC) from a treated effluent, along with the subsequent reusability of the NC. Several research groups have begun to focus on magnetically separable photocatalysts for wastewater treatment, demonstrating the value of the special properties of magnetic materials. Among these magnetic materials, Fe 2 O 3 has a narrow bandgap, high chemical resistance, and high resistance to corrosion. Therefore, rationally designed MoS 2 /Fe 2 O 3 NCs can serve as a reusable SERS substrate for detection and easily reclaimed photocatalyst. The recovery and economical reuse of MoS 2 /Fe 2 O 3 NCs photocatalysts is easily achieved by adding an external magnetic field.
Bisphenol A (BPA) is believed to be an endocrine disruptor and widely exists in food containers and the environment. Even low levels of BPA entering the body can disrupt the endocrine system by binding to estrogen receptors, which may lead to cardiovascular diseases, immune function deficiencies, impaired reproductive capacity, and other diseases [21][22][23] . Thus, it is imperative to develop a facile, rapid, and inexpensive method for BPA detection and degradation.
In this study, MoS 2 /Fe 2 O 3 NCs were prepared via a simple low-temperature hydrothermal method, and the advantages of the two materials were combined. For example, after 50 min of ultraviolet (UV) irradiation, the substrate completely degraded BPA, and upon recovery, demonstrated its detection capability. Compared with MoS 2 NFs and Fe 2 O 3 NPs, the rate of degradation of BPA and the SERS activity of MoS 2 /Fe 2 O 3 NCs were significantly better. This new, easily recoverable SERS sensor with a high sensitivity will facilitate sensing harmful molecules. To the best of our knowledge, no MoS 2 /Fe 2 O 3 composites that exhibit BPA detection and photocatalysis multifunctionality have been reported thus far. Photocatalytic and SERS mechanisms were also proposed.

Characterization analysis of MoS 2 /Fe 2 O 3 NCs
A growth flow diagram of the MoS 2 /Fe 2 O 3 NCs is shown in Fig. 1. Figure 2a confirms that the MoS 2 sample was pure hexagonal 2H-MoS 2 (JCPDS card no. 37-1492). The peak with the highest intensity (at 2θ = 14.09°) indicated that MoS 2 had excellent lamellar growth in the c-axis direction. In regard to MoS 2 /Fe 2 O 3 , some of the peaks corresponded to 2H-MoS 2 , while others corresponded to tetragonal γ-Fe 2 O 3 (JCPDS card no. 39-1346) phase, indicating that the native structure of each constituent was well preserved during the reaction. The intensities of the MoS 2 peaks for the MoS 2 /Fe 2 O 3 NCs were lower than those for pure MoS 2 because the Fe 2 O 3 NPs attached to the MoS 2 nanoflowers (NFs). Raman spectra confirmed the chemical composition of the MoS 2 /Fe 2 O 3 NCs and MoS 2 . Two characteristic Raman peaks of MoS 2 were observed at 337 and 377 cm −1 , corresponding to the A 1g and 1 E 2g vibration modes, respectively; additionally, their peak frequency difference was Δk = 40 cm −1 (Fig. 2b) [24][25][26] . However, after the incorporation of Fe 2 O 3 , the characteristic Raman peaks of MoS 2 shifted to 338 and 379 cm −1 , and the peak frequency difference was Δk = 41 cm −1 . Δk represents the number of . In regard to the MoS 2 /Fe 2 O 3 NCs, the lattice spacing was 0.624 nm, corresponding to the hexagonal MoS 2 (0 0 2) plane. In addition, the boundary between Fe 2 O 3 and MoS 2 was clearly observed, indicating that a heterojunction was formed between these two components. The regions with different colors in Fig. 2j-m correspond to S, Mo, Fe, and O, and the elemental distribution in MoS 2 /Fe 2 O 3 was uniform.
X-ray photoelectron spectroscopy (XPS) was performed to analyze the electronic states and chemical composition of the MoS 2 /Fe 2 O 3 NCs (Fig. 3). The survey scan spectra of pristine MoS 2 , Fe 2 O 3 , and MoS 2 /Fe 2 O 3 NCs are presented in Fig. 3a, which confirmed the coexistence of Fe 2p, O 1s, Mo 3p, and S 2p in the hybrid. The Mo 3d spectra exhibited three peaks for pristine MoS 2 , but after forming the MoS 2 /Fe 2 O 3 NCs, four peaks appeared in Fig. 3b. The peaks at 235.8, 232.6, 229.4, and 226.5 eV corresponded to Mo 6+ 3d 3/2 , Mo 4+ 3d 3/2 , Mo 4+ 3d 5/2 , and S 2s, respectively. A small portion of Mo 4+ was oxidized into Mo 6+ during the reaction, confirming that

SERS enhancement and reusability of MoS 2 /Fe 2 O 3 NCs for BPA detection
Herein, p-aminobenzene sulfonic acid, NaNO 3 , and Na 2 CO 3 as Pauly's reagent were added in the BPA detection test to enhance the adhesion of BPA on the surface of the MoS 2 /Fe 2 O 3 NCs. As shown in Fig. 4, p-aminobenzene sulfonic acid, NaNO 3 and Na 2 CO 3 , were all low Raman scattering active molecules; therefore, their addition had almost no effect on BPA detection. To confirm that MoS 2 /Fe 2 O 3 had excellent SERS properties, SERS spectra of BPA absorbed on MoS 2 /Fe 2 O 3 at various concentrations ranging from 10 −4 to 10 −9 M were obtained, as shown in Fig. 5a. These results indicated that as the concentration of BPA decreased, the intensity of the Raman peaks decreased. The characteristic peak of BPA at 1124 cm −1 was observed at concentrations as low as 10 −9 M, indicating that the MoS 2 /Fe 2 O 3 NCs had a high sensitivity. The intensity of the peak at 1124 cm −1 was correlated with the BPA concentration; thus, we used it for further quantitative analysis. Figure 5b shows the direct proportionality between the BPA concentration, in the range of 10 −4 -10 −9 M, and the normalized Raman signal intensity. The linear equation is as follows: with a squared correlation coefficient of R 2 = 0.97. The stability of the substrate is an important factor that must be considered. As shown in Fig. 5c    Detection in "real-world" samples To evaluate the application of the MoS 2 /Fe 2 O 3 NCs, "real-world" samples (lake water and milk) were chosen for detection. As shown in Fig. 6, the characteristic CH wagging peak of BPA at 1124 cm −1 was observed at concentrations as low as 10 −7 M for these samples, indicating that the MoS 2 /Fe 2 O 3 NCs could be used for the practical and rapid detection of BPA.

Photocatalytic activity of MoS 2 /Fe 2 O 3 NCs
The catalytic properties of the pristine MoS 2 Fig. 7a-c, respectively. For all catalysts, the intensity of the main absorption peak decreased with increasing irradiation time. After 50 min of UV irradiation, the degradation rates of the two pristine photocatalysts (MoS 2 NFs and Fe 2 O 3 NPs) were only~40% and 48%, respectively. Surprisingly, the photocatalytic activity was significantly increased in the presence of the MoS 2 /Fe 2 O 3 NCs catalyst; in this case, >92% of the present BPA was decomposed after 50 min of irradiation, as shown in Fig. 7c. This degradation rate is significantly higher than those observed with the MoS 2 NFs and Fe 2 O 3 NPs.
A related graph showing the dependence of the BPA degradation efficiencies of the MoS 2 /Fe 2 O 3 NCs and other catalysts on the UV irradiation time is presented in Fig. 8a. We define the degradation efficiency as C/C 0 , where C 0 represents the initial BPA concentration (mg/L) and C represents the BPA concentration after the reaction (mg/L). As shown in Fig. 8a, the MoS 2 /Fe 2 O 3 NCs had better photocatalytic activity than the other catalysts. The photocatalytic efficiency of MoS 2 /Fe 2 O 3 NCs was as high as 0.02, which was higher than that of pure MoS 2 (0.01) and Fe 2 O 3 (0.008). Thus, the MoS 2 / Fe 2 O 3 NCs has great potential for use in wastewater treatments. Before light irradiation, the photocatalyst and BPA solution were stirred under dark conditions for 10 min to attain an adsorption equilibrium. During this period, the concentration of BPA decreased because of the adsorption of BPA molecules on the photocatalysts. We used the pseudo-first-order mode to investigate the reaction kinetics of BPA degradation. The simplified equation is: where k represents the apparent first-order reaction rate constant 35 . Figure 8b  between −ln(C/C 0 ) and the irradiation time for different photocatalysts. The curves could be fitted with a linear relationship, indicating that the degradation kinetics followed a typical first-order reaction. Using Eq. (1), we determined the apparent pseudo-firstorder rate constants for the different photocatalysts.
The k values of the pristine Fe 2 O 3 NPs, MoS 2 NFs, and MoS 2 /Fe 2 O 3 NCs were calculated to be 0.69, 0.53, and 2.41, respectively. Stable photoactivity under UV light is critical for practical water treatment applications, particularly for composite materials that may lose their coating. We examined the loss of the BPA degradation activity of the MoS 2 /Fe 2 O 3 NCs by utilizing it for five consecutive cycles under UV light irradiation. No loss of activity was observed (Fig. 8c). As shown in Fig. 8d, the structure of the catalyst was not significantly changed after five consecutive photocatalytic degradation cycles, also suggesting that the Fe 2 O 3 nanoparticles could slow down the photocorrosion of MoS 2 , thereby efficiently protecting MoS 2 . Generally, MoS 2 is prone to photocorrosion due to oxidation of surface sulfions to sulfurs by photoexcited holes. Therefore, the MoS 2 / Fe 2 O 3 NCs exhibited high stability and excellent anti-photocorrosion properties, showing that this material has promise for use in environmental restoration applications.

Mechanisms of SERS detection and photocatalysis
When the MoS 2 /Fe 2 O 3 heterojunction system was irradiated with UV light, MoS 2 was excited, generating electron-hole pairs because of its narrow bandgap. The photoinduced electrons moved rapidly from the conduction band (CB) of MoS 2 to that of Fe 2 O 3 , as shown in Fig. 9. In the MoS 2 /Fe 2 O 3 NCs, the spatial separation of photoexcited holes and electrons extended the chargecarrier lifetime and hindered the recombination of electron-hole pairs, thereby enhancing the photocatalytic activity. Moreover, the selected transfer of holes from the valence band (VB) of MoS 2 to Fe 2 O 3 remarkably weakened the photocorrosion activity. After the carriers of MoS 2 and Fe 2 O 3 were generated, the free electrons accumulated in the CB of Fe 2 O 3 , while photoinduced holes were present in the VB of MoS 2 ; thus, a high photocatalytic activity was obtained. Effective Z-type electron-hole pair separation and an effective transfer path were achieved, and a strong redox capacity of the photoexcited electron and holes was obtained in the CB and VB, respectively, significantly improving the photocatalytic and SERS activity of the MoS 2 /Fe 2 O 3 NC heterojunction. Therefore, Fe 2 O 3 not only acted as a protective shell for the MoS 2 core by preventing the loss of sulfur but also constructed Z-type junctions that prolonged the lives of photogenerated electrons and holes, which would significantly enhance the photocatalytic activity and stability. Another reason for the SERS enhancement was the semiconducting nature of MoS 2 . Because its surface had S atoms and polar covalent bonds (Mo-S) perpendicular to the surface, this dipole-dipole coupling significantly increased the intensity of the Raman peaks 36 . In addition, because of the large surfaceto-volume ratio, there was an abundance of active adsorption sites for gas molecules. The reactions involved in the photocatalytic process are summarized as follows:

Conclusion
In summary, a multifunctional material was fabricated by simply depositing Fe 2 O 3 NPs onto MoS 2 NFs, which significantly improved its photocatalytic properties and ability to be used as a SERS substrate. In addition, the MoS 2 /Fe 2 O 3 NCs were successfully recycled. This study is the first to report MoS 2 /Fe 2 O 3 NCs used as SERS substrates for BPA detection. The MoS 2 /Fe 2 O 3 NCs had a detection limit of 1 × 10 −9 M, along with exhibiting excellent stability. The prepared MoS 2 /Fe 2 O 3 NCs had higher photocatalytic activity than the MoS 2 NFs and Fe 2 O 3 NPs alone. The enhanced photocatalytic activity and SERS activity were attributed to the efficient separation and transfer of electron-hole pairs by the Z-scheme heterojunction system. Therefore, as efficient multifunctional catalysts, MoS 2 / Fe 2 O 3 NCs are expected to not only replace metal catalysts for removing organic matter from water and the environment but also pave the way for SERS applications, thereby introducing new methods for chemical and medical detection and for environmental monitoring.

MoS2 NF preparation
First, 0.5 g H 4 MoNa 2 O 6 and 0.7 g CH 4 N 2 S were mixed and stirred in 70 mL of ultrapure water. Then, 0.5 g C 6 H 8 O 7 ·H 2 O was added until complete dissolution was achieved. The sample was transferred into an 80 mL Teflon-lined hydrothermal autoclave reactor and then placed in a drying box at 240°C for 24 h. Next, the reaction products were centrifuged with alcohol and ultrapure water and dried at 70°C.

MoS 2 /Fe 2 O 3 NC preparation
First, 0.2 g of MoS 2 powder, 0.5 g of H 18 FeN 3 O 18 , and 0.7 g of H 2 NCONH 2 were mixed in 70 mL of ultrapure water. Then, 0.02 g of C 18 H 29 NaO 3 S were well dispersed in the liquid mixture, stirred in a 60°C water bath for 35 min, transferred to an 80 mL reactor, and finally placed in a drying box at 90°C for 12 h. The MoS 2 /Fe 2 O 3 NCs were washed with absolute ethanol and water to remove possible residuals. The solid powder solid was placed in a drying box and kept dry at 80°C.

Characterizations
XPS (ESCALAB250X, Thermo Scientific) and X-ray diffraction (XRD, D/Max 3C, Rigaku) were used to study the structural quality. TEM (JEM-2100HR, JEOL) and SEM (JSM-7800F, JEOL) were used to characterize the morphology of the samples. UV-visible absorption spectroscopy (UV-3600, Shimadzu Corporation) and a vibrating sample magnetometer (7407, Lake Shore) were used to characterize the optical and magnetic properties of the samples. Raman spectra were obtained with an Ar + -ion laser (inVia Raman, Renishaw).

SERS experiments of BPA
We used the coupling reaction of BPA with Pauly's reagents (p-aminobenzene sulfonic acid, HCl, NaNO 3 , and Na 2 CO 3 ) to enhance the adhesion of BPA onto the surface of the SERS substrate materials. Please refer to our previous report for the detailed process 37 .

SERS experiments of BPA in milk and lake water
Real milk contains fat, protein, vitamins, and other organic ingredients that can interfere with the detection of BPA. Therefore, it is necessary to pretreat the milk sample with BPA. The process is as follows. First, methanol (7 mL) and water (3 mL) were mixed and added to the milk sample (containing 10 −7 M BPA, 4 mL), and then the mixture was sonicated and centrifuged at 10,000 r.p.m. for 3 min. The upper supernatant was extracted and then dried. This extract was collected in another centrifuge tube and mixed with methanol and water, with the above sonication and centrifugation process being repeated. Finally, the extract was filtered by membrane filters (0.45 and 0.22 µm) for the SERS test. The procedures for the detection and data analysis were the same as those for detecting BPA in water.
We collected lake water from a local source (South Lake in Changchun City). Lake water samples with BPA added were filtered by membrane filters (0.45 and 0.22 µm) before the detection test to avoid interference from other impurities. The test process was consistent with that described above.

Photodegradation experiments
In the degradation process, circulating water was used to ensure that all the tests were performed at room temperature. The photodegradation of BPA under UV light was performed to assess the activity of photocatalysts. One hundred milliliters of an aqueous solution was prepared with 0.001 g of BPA and 0.05 g of Fe 2 O 3 / MoS 2 , MoS 2 , or Fe 2 O 3 NCs. The test solution was stirred magnetically in a 100-mL beaker. The sample was kept in a dark room for 10 min for the adsorption of BPA molecules on the photocatalysts before being subjected to UV irradiation. During the experiment, the samples were taken at specified times. After each sampling, the catalyst was separated via centrifugation for testing.