Extended Near-Infrared Photoactivity of Bi6Fe1.9Co0.1Ti3O18 by Upconversion Nanoparticles

Bi6Fe1.9Co0.1Ti3O18 (BFCTO)/NaGdF4:Yb3+, Er3+ (NGF) nanohybrids were successively synthesized by the hydrothermal process followed by anassembly method, and BFCTO-1.0/NGF nanosheets, BFCTO-1.5/NGF nanoplates and BFCTO-2.0/NGF truncated tetragonal bipyramids were obtained when 1.0, 1.5 and 2.0 M NaOH were adopted, respectively. Under the irradiation of 980 nm light, all the BFCTO samples exhibited no activity in degrading Rhodamine B (RhB). In contrast, with the loading of NGF upconversion nanoparticles, all the BFCTO/NGF samples exhibited extended near-infrared photoactivity, with BFCTO-1.5/NGF showing the best photocatalytic activity, which could be attributed to the effect of {001} and {117} crystal facets with the optimal ratio. In addition, the ferromagnetic properties of the BFCTO/NGF samples indicated their potential as novel, recyclable and efficient near-infrared (NIR) light-driven photocatalysts.

Recently, layered bismuth-based Bi 4 Bi n−3 Fe n−3−x Co x Ti 3 O 3n+3 (BFCTO) Aurivillius compounds, which are formed by perovskite-type (Bi n−1 Fe n−3−x Co x Ti 3 O 3n+1 ) 2− blocks sandwiched between fluorite-type (Bi 2 O 2 ) 2+ slabs, where n is the number of perovskite units per half-cell [18][19][20][21], proved to be apotential single-phase multiferroic material with ferroelectricity and ferromagnetism [22]. Moreover, the internal electric field and ferromagnetism could enable the photocatalysts recyclable in viscous solution, which is important but extremely difficult due to their size and high viscosity [23]. In 2008, nanostructured Bi 5 FeTi 3 O 15 was first realized, and showed excellent visible light-driven photocatalytic activity compared to bulk-Bi 5 FeTi 3 O 15 and the widely-used photocatalyst TiO 2 [24]. In 2014, Mandal et al. [25] successfully synthesized La 3+ doped Bi 5 Ti 3 FeO 15 (n = 4) via the solid-state reaction, and found that such layered oxides exhibited efficient photocatalysis for Rhodamine B (RhB) degradation under sunlight irradiation. In a previous study, the authors [26] optimized the photocatalysis in ferromagnetic Bi 6 Fe 1.9 Co 0.1 Ti 3 O 18 (n = 5) nanocrystals by morphology control, and found that the synergistic effect of {001} and {117} facets in the separation of electrons and holes, and the oxidation/reduction reaction, efficiently inhibited the recombination of the charge carrier. Furthermore, we found the appropriate area ratio between the {001} facets and the {117} facets was shown to be beneficial to photocatalysis efficiency. Recently, we also obtained visible light-responsive Bi 7 Fe 3 Ti 3 O 21 nanoshelf photocatalysts [27]. However, BFCTO photocatalysts are still limited to ultraviolet (UV) and/or visible (Vis) light, while near-infrared (NIR) light driven BFCTO-based catalysts have never been reported, despite NIR light making up about 47% of the solar spectrum. Therefore, the extended utilization of sunlight from Vis to NIR light is an attractive issue for developing photocatalysts. On the other hand, upconversion materials could convert low-energy NIR light into high-energy UV or Vis light through a nonlinear optical process [28][29][30][31][32][33]. Furthermore, compared with conventional fluorescent materials, such as organic dyes and quantum dots, upconversion nanoparticles have the following advantages: (i) High photostability and chemical stability; (ii) low toxicity to the human body; (iii) good biocompatibility by surface modification and functionalization; and (iv) good optical transparency over a wide wavelength range and low phonon energy [34][35][36][37]. Therein, coupling upconversion nanoparticles with photocatalysts is considered to bea very promising method to extend NIR photoactivity [38].
In this article, we synthesized firstly Bi 6 Fe 1.9 Co 0.1 Ti 3 O 18 (BFCTO)/NaGdF 4 :Yb 3+ , Er 3+ (NGF) nanohybrids by the hydrothermal process, followed by an assembly method, where the morphology and size of the BFCTO component was controlled by adjusting NaOH concentration. The structures, morphologies, upconversion emission, UV-Vis-NIR diffuse reflectance spectra and magnetic properties of the samples were characterized, and the photocatalytic activities were also evaluated by the degradation of RhB solution under the irradiation of light of wavelength 980 nm. The loading of the NGF upconversion nanoparticles extended the near-infrared photoactivity of all the BFCTO samples. BFCTO-1.5/NGF exhibited the best photocatalytic activity, and thus the corresponding mechanism was investigated in detail.

Materials
All chemicals were purchased from Sinopharm Chemical Reagent Co., Ltd., (Shanghai, China), without further purification. The details are as follows: Ti(OC 4 H 9 ) 4

Synthesis of BFCTONanoparticles with Various Morphologies
The BFCTO nanoparticles were synthesized by the hydrothermal method according to our previous report [23], and the corresponding morphologies were controlled by adjusting NaOH concentrations. Here, BFCTO-1.0, BFCTO-1.5 and BFCTO-2.0 represent BFCTO samples obtained with 1.0 M, 1.5 M and 2.0 M NaOH, respectively.

Preparation of PVP-Capped BFCTO Nanoparticles
The as-prepared BFCTO nanoparticles were added into the PVP solution (1.0 g PVP dissolved in 100 mL deionized water), and stirred for 12 h at room temperature. The obtained samples were

Synthesis of Oleic-Capped and Free NGF Nanoparticles
The oleic-capped NGF nanoparticles were synthesized using the solvothermal method. Firstly, 10 mL oleic acid and 3 mL NaOH solution (5 M) were successively added into 10 mL ethanol. After stirring for 10 min, 3.12 mL Gd(NO 3 ) 3 , 0.80 mL Yb(NO 3 ) 3 , 0.08 mL Er(NO 3 ) 3 and 2 mL NH 4 F (2 M) were added. Then the solution was transferred into a 50-mL Teflon-lined stainless-steel autoclave. Theautoclave was sealed and heated at 200 • C for 3 h and then cooled to room temperature. The obtained nanoparticles were collected by centrifugation, and then washed with water and ethanol for several cycles. Then the as-prepared oleic acid-capped NGF nanoparticles were dispersed in HCl solution (0.1 M) and ultrasonicated for 15 min to remove the ligands from the surface of the nanoparticles. The obtained NGF samples were collected via centrifugation and further purified by ethanol and deionized water for several rounds, and finally re-dispersed in deionized water.

Synthesis of BFCTO/NGF Samples
The as-prepared 0.3 g PVP-capped BFCTO nanoparticles were dissolved in 100 mL H 2 O, and 10 mL NGF solution (0.01 g mL −1 ) was added. Then, the solution was stirred for 12 h. The obtained nanoparticles were collected by centrifugation and washed several times with deionized water, and dried at 80 • C under vacuum for 12 h.

Photocatalytic Activities
Photocatalytic activities of the as-prepared BFCTO/NGF samples were evaluated by the photodecomposition of RhB under the NIR irradiation of 980 nm (power = 1.0 A). Firstly, 50 mg of BFCTO/NGF and the BFCTO sampleswere dispersed uniformly into a 50-mL RhB solution (5 mg L −1 ). Then, the suspension was stirred for 30 min in the dark to ensure the establishment of an adsorption/desorption equilibrium. During the NIR irradiation, 3 mL reaction solution was taken and centrifuged every half an hour, and the filtrate was measured on a UV-Visible spectrometer at a maximum absorption wavelength of 554 nm to determine the concentration of RhB.

Results and Discussion
The detailed procedures of the BFCTO/NGF nanocomposites are outlined in Figure 1 and the Experimental Section. Figure 2 displays the XRD patterns of the NGF, BFCTO and BFCTO/NGF samples. Strong and sharp diffraction peaks were observed in all the samples, suggesting that the compounds are well-crystallized. The XRD pattern recorded in red suggests the formation of BFCTO pure phase, while all the XRD patterns indicated the characteristic BFCTO peaks with an orthorhombic structure (with a space group of A2 1 am, PDF 38-1257) [39]. The pattern recorded in pink reveals that the NGF hexagonal phase is obtained according to PDF 27-0699. The XRD pattern in blue displays both the characteristic peaks of the BFCTO and NGF phases, which suggests that NGF may be loaded onto BFCTO successfully. orthorhombic structure (with a space group of A21am, PDF 38-1257) [39].The pattern recorded in pink reveals that the NGF hexagonal phase is obtained according to PDF 27-0699. The XRD pattern in blue displays both the characteristic peaks of the BFCTO and NGF phases, which suggests that NGF may be loaded onto BFCTO successfully. The representative SEM images of the BFCTO-1.5 and BFCTO-1.5/NGF samples are shown in Figure 3a-c. BFCTO-1.5 nanoparticles are monodisperse and exhibit a nanoplate morphology, with a thickness and edge length of ~100 nm and ~1μm, respectively ( Figure 3a).There are two key factors for enhancing the catalytic activity of BFCTO by upconversion nanoparticles: (i) PVP surfactant could coordinate lanthanide ions through pyrrolidone groups on the surface of the NGF nanoparticles, which ensures the successful coating of NGF on the surface of BFCTO [40]; and(ii) the oleate ligand capped on the NGF nanoparticles is protonated by hydrochloric acid, thereby releasing oleic acid (OA) from the surface. This release of OA may increase the water solubility and reduce the fluorescence quenching of the NGF upconversion nanoparticles [41,42], which is beneficial for enhancing the NIR-responsive photocatalysis of BFCTO by upconversion nanoparticles. Figure 3b-d validates the successful loading of the NGF nanoparticles onto the BFCTO-1.5 nanoplates due to the modification of the PVP surfactant, with the size of the NGF nanoparticlesbeing ~30nm (Figure 3e).Furthermore, the characteristic peaks of the oleate ligand disappear in the FTIR spectrum of the NGF nanoparticles, as shown Figure S1, suggesting that the NGF nanoparticles are oleic-free. HRTEM was also adopted to characterize the microstructures of orthorhombic structure (with a space group of A21am, PDF 38-1257) [39].The pattern recorded in pink reveals that the NGF hexagonal phase is obtained according to PDF 27-0699. The XRD pattern in blue displays both the characteristic peaks of the BFCTO and NGF phases, which suggests that NGF may be loaded onto BFCTO successfully. The representative SEM images of the BFCTO-1.5 and BFCTO-1.5/NGF samples are shown in Figure 3a-c. BFCTO-1.5 nanoparticles are monodisperse and exhibit a nanoplate morphology, with a thickness and edge length of ~100 nm and ~1μm, respectively (Figure 3a).There are two key factors for enhancing the catalytic activity of BFCTO by upconversion nanoparticles: (i) PVP surfactant could coordinate lanthanide ions through pyrrolidone groups on the surface of the NGF nanoparticles, which ensures the successful coating of NGF on the surface of BFCTO [40]; and(ii) the oleate ligand capped on the NGF nanoparticles is protonated by hydrochloric acid, thereby releasing oleic acid (OA) from the surface. This release of OA may increase the water solubility and reduce the fluorescence quenching of the NGF upconversion nanoparticles [41,42], which is beneficial for enhancing the NIR-responsive photocatalysis of BFCTO by upconversion nanoparticles. Figure 3b-d validates the successful loading of the NGF nanoparticles onto the BFCTO-1.5 nanoplates due to the modification of the PVP surfactant, with the size of the NGF nanoparticlesbeing ~30nm (Figure 3e).Furthermore, the characteristic peaks of the oleate ligand disappear in the FTIR spectrum of the NGF nanoparticles, as shown Figure S1, suggesting that the The representative SEM images of the BFCTO-1.5 and BFCTO-1.5/NGF samples are shown in Figure 3a-c. BFCTO-1.5 nanoparticles are monodisperse and exhibit a nanoplate morphology, with a thickness and edge length of~100 nm and~1 µm, respectively (Figure 3a). There are two key factors for enhancing the catalytic activity of BFCTO by upconversion nanoparticles: (i) PVP surfactant could coordinate lanthanide ions through pyrrolidone groups on the surface of the NGF nanoparticles, which ensures the successful coating of NGF on the surface of BFCTO [40]; and (ii) the oleate ligand capped on the NGF nanoparticles is protonated by hydrochloric acid, thereby releasing oleic acid (OA) from the surface. This release of OA may increase the water solubility and reduce the fluorescence quenching of the NGF upconversion nanoparticles [41,42], which is beneficial for enhancing the NIR-responsive photocatalysis of BFCTO by upconversion nanoparticles. Figure 3b-d validates the successful loading of the NGF nanoparticles onto the BFCTO-1.5 nanoplates due to the modification of the PVP surfactant, with the size of the NGF nanoparticlesbeing~30 nm (Figure 3e). Furthermore, the characteristic peaks of the oleate ligand disappear in the FTIR spectrum of the NGF nanoparticles, as shown Figure S1, suggesting that the NGF nanoparticles are oleic-free. HRTEM was also adopted to characterize the microstructures of the NGF nanoparticles, as shown in the insert of Figure 3d, and it indicated that the spacing between two adjacent lattice planes was 0.30 nm, fitting well with the (110) plane of the hexagonal phase of NGF (PDF 27-0699). To further verify the successful synthesis of the BFCTO-1.5/NGF nanocomposites, HADDF-STEM imaging and the EDS elemental mapping were conducted, as shown in Figure 3f,g. The HADDF-STEM image shows a sharp contrast between the NGF and BFCTO-1.50 nanoparticles, consistent with the TEM images in Figure 3b,c. The EDS elemental mapping in Figure 3g indicated that Na, Gd and F elements were distributed in the shell, while Bi, Fe, Co, Ti and O existed in the core, which further demonstrated the successful synthesis of BFCTO-1.50/NGF nanohybrids.  Additionally, XPS spectra were also conducted to determine the valence states of metal elements in the BFCTO/NGF nanohybrids (Figure 4). The bindingenergies obtained from the XPS analyses were corrected for specimen charging by referencing the C1s line to 284.5 eV. The full survey spectrum revealed that theBFCTO-1.5/NGF nanohybrid was composed of Bi, Fe, Co, Ti, O, Na, Gd and F elements (Figure 4a). Figure 4b illustrates the core-level spectrum of Bi4f in the nanohybrid, and the peaks located at 158.7 eV (Bi 4f7/2) and 164.0 eV (Bi 4f5/2) are referenced by Bi2O3 [43]. Figure 4c shows the Fe 2p core-level spectrum of the BFCTO-1.5/NGF nanohybrid, and the two peaks of Fe 2p at 723.81 and 710.2 eV-which match well with the spin-orbit split of Fe 2p1/2 and Fe 2p3/2, respectively [39]-are near to those in Fe2O3, suggesting that the Fe ions in the BFCTO/NGF nanocomposite have +3 valence states. The binding energies at 465.15 and 457.26 eV correspond to the binding energy of Ti 2p1/2 and Ti 2p3/2 (Figure 4d), implying that the doped Ti ions in the BFCTO/NGF nanohybrid have +4 valence states [44]. Additionally, Figure 4e shows the Gd Additionally, XPS spectra were also conducted to determine the valence states of metal elements in the BFCTO/NGF nanohybrids (Figure 4). The bindingenergies obtained from the XPS analyses were corrected for specimen charging by referencing the C 1s line to 284.5 eV. The full survey spectrum revealed that the BFCTO-1.5/NGF nanohybrid was composed of Bi, Fe, Co, Ti, O, Na, Gd and F elements (Figure 4a). Figure 4b illustrates the core-level spectrum of Bi 4f in the nanohybrid, and the peaks located at 158.7 eV (Bi 4f 7/2 ) and 164.0 eV (Bi 4f 5/2 ) are referenced by Bi 2 O 3 [43]. Figure 4c shows the Fe 2p core-level spectrum of the BFCTO-1.5/NGF nanohybrid, and the two peaks of Fe 2p at 723.81 and 710.2 eV-which match well with the spin-orbit split of Fe 2p 1/2 and Fe 2p 3/2 , respectively [39]-are near to those in Fe 2 O 3 , suggesting that the Fe ions in the BFCTO/NGF nanocomposite have +3 valence states. The binding energies at 465.15 and 457.26 eV correspond to the binding energy of Ti 2p 1/2 and Ti 2p 3/2 (Figure 4d), implying that the doped Ti ions in the BFCTO/NGF nanohybrid have +4 valence states [44]. Additionally, Figure 4e shows the Gd 4d 3/2 peak at 149.7 eV and the Gd 4d 5/2 peak at 143.9 eV [45], and the Gd 3d 5/2 peak at 1188.5 eV and Gd 3d 7/2 peak at 1220.9 eV are also observed (Figure 4e) [46], which are characteristic of Gd 3+ ions. Therefore, the XPS results indicate that the metal elements exist in the form of Bi 3+ , Fe 3+ , Ti 4+ and Gd 3+ . and edge length of ~40 nm and ~1μm, respectively (Figure 5a). The BFCTO-2.0 sample exhibits a well-defined shape of a truncated tetragonal bipyramid, the thickness and edge length of which ranges from 100 nm to 1 μm and 1 to 2 μm, respectively (Figure 5c) [26]. Evidently, the coating of NGF did not change the size and morphology of BFCTO because the NGF nanoparticles were just absorbed onto the BFCTO nanoparticles' surfaces through the amide bond in PVP.  After successfully synthesizing the BFCTO/NGF nanoparticles, we were in a position to evaluate the corresponding photocatalytic activities. To identify the dependence of morphology on photocatalysis performance, the degradation rate of all the samples was normalized by the specific surface area, which is 17.733, 11.731 and 11.610 m 2 g −1 for BFCTO-1.0/NGF, BFCTO-1.5/NGF and BFCTO-2.0/NGF, respectively. Figure Figure 5. The BFCTO-1.0 sample displays a nanosheet-like morphology, with a thickness and edge length of~40 nm and~1 µm, respectively (Figure 5a). The BFCTO-2.0 sample exhibits a well-defined shape of a truncated tetragonal bipyramid, the thickness and edge length of which ranges from 100 nm to 1 µm and 1 to 2 µm, respectively (Figure 5c) [26]. Evidently, the coating of NGF did not change the size and morphology of BFCTO because the NGF nanoparticles were just absorbed onto the BFCTO nanoparticles' surfaces through the amide bond in PVP.
After successfully synthesizing the BFCTO/NGF nanoparticles, we were in a position to evaluate the corresponding photocatalytic activities. To identify the dependence of morphology on photocatalysis performance, the degradation rate of all the samples was normalized by the specific surface area, which is 17.733, 11.731 and 11.610 m 2 g −1 for BFCTO-1.0/NGF, BFCTO-1.5/NGF and BFCTO-2.0/NGF, respectively. Figure 6a    After successfully synthesizing the BFCTO/NGF nanoparticles, we were in a position to evaluate the corresponding photocatalytic activities. To identify the dependence of morphology on photocatalysis performance, the degradation rate of all the samples was normalized by the specific surface area, which is 17.733, 11.731 and 11.610 m 2 g −1 for BFCTO-1.0/NGF, BFCTO-1.5/NGF and BFCTO-2.0/NGF, respectively. Figure 6a displays the irradiation time dependence of RhB degradations in various BFCTO and BFCTO/NGF aqueous dispersions under NIR irradiation of 980nm, with a power of 1.0A. Evidently, all the BFCTO samples exhibited no activity after the adsorption/desorption equilibrium. In contrast, with the loading of NGF upconversion nanoparticles, NIR-responsive photocatalysis activity was observed in all the BFCTO/NGF samples,  To explain the effect of NGF-loading on BFCTO nanoparticles with different morphologies, the upconversion emission spectra of all the BFCTO/NGF samples were measured (Figure 6b). Under the excitation of a980nm continuous wave diode laser, the green emission centered at 543 nm and red emission centered at 658 nm were observed, originating from the transitions 2 H11/2/ 4 S3/2 → 4 I15/2 and 4 F9/2→ 4 I15/2, respectively, for theEr 3+ ions, indicating that the loading of NGF nanoparticles could extend the absorption of BFCTO from Vis to NIR light. All of the BFCTO/NGF samples exhibited stronger red emission than green emission due to high Yb 3+ doping concentration (20 at%), and stronger absorption of BFCTO in the green band (Figure 6c). Furthermore, the intensity of the upconversion emissions decreased in the order of IBFCTO-2.0/NGF<IBFCTO-1.5/NGF<IBFCTO-1.0/NGF (Figure 6b), suggesting that absorption in the NIR region increases in the order of To explain the effect of NGF-loading on BFCTO nanoparticles with different morphologies, the upconversion emission spectra of all the BFCTO/NGF samples were measured (Figure 6b). Under the excitation of a 980 nm continuous wave diode laser, the green emission centered at 543 nm and red emission centered at 658 nm were observed, originating from the transitions 2 H 11/2 / 4 S 3/2 → 4 I 15/2 and 4 F 9/2 → 4 I 15/2 , respectively, for the Er 3+ ions, indicating that the loading of NGF nanoparticles could extend the absorption of BFCTO from Vis to NIR light. All of the BFCTO/NGF samples exhibited stronger red emission than green emission due to high Yb 3+ doping concentration (20 at%), and stronger absorption of BFCTO in the green band (Figure 6c). Furthermore, the intensity of the upconversion emissions decreased in the order of I BFCTO-2.0/NGF < I BFCTO-1.5/NGF < I BFCTO-1.0/NGF (Figure 6b), suggesting that absorption in the NIR region increases in the order of I BFCTO-2.0/NGF > I BFCTO-1.5/NGF > I BFCTO-1.0/NGF , consistent with the results in Figure 6c. Furthermore, upconversion nanoparticles may reduce the electrical conductivity of BFCTO photocatalysts due to NGF's electrical insulating properties. The UV-Vis-NIR diffuse reflectance spectra in the wavelength range of 250-1500 nm for all the BFCTO/NGF samples, displayed in Figure 6c, are relevant to the electronic structure, and are the key factors to determine the band gaps. According to the UV-Vis-NIR diffuse reflectance spectra in Figure 6c, the band gap values are calculated by the following equation: αhv = A (hv − E g ) n/2 , where α, hv, A, and E g represent the absorption coefficient, photon energy, proportionality constant and gap band, respectively. BFCTO possesses a direct electron transition, therein n = 1. Figure 6d shows the plot of the (αhv) 2 verse E g , from which the band gaps of the BFCTO/NGF samples were estimated by extrapolating a straight line to the abscissa axis. The band gaps were found to be 2.48, 2.47 and 2.43 eV for BFCTO-1.0/NGF, BFCTO-1.5/NGF and BFCTO-2.0/NGF, respectively. The intrinsic absorption edge of BFCTO-2.0/NGF has an obvious red-shift by 10 nm to that of BFCTO-1.0/NGF.
The photodegradation mechanism of BFCTO/NGF under 980 nm irradiation is shown in Figure 7. In the upconversion luminescent NGF nanoparticle, the excited Yb 3+ ions act as sensitizers to absorb the 980 nm photons, and then transfer this energy to Er 3+ ions to emit Vis light. The emission peaks centered at 524, 543, and 658 nm for NGF are attributed to the Er 3+ transitions of 2 H 11/2 → 4 I 15/2 , 4 S 3/2 → 4 I 15/2 , and 4 F 9/2 → 4 I 15/2 , respectively. Meanwhile, these emission peaks are in the Vis range, which are fully in the BFCTO absorption range, leading to photoexcitation of the BFCTO nanoparticles in BFCTO/NGF. Therefore, the holes (h + ) on the valence band (VB) and the electrons (e − ) on the conduction band (CB) were able to be generated to decompose the RhB molecules. The oxidative holes could react with H 2 O to produce hydroxyl radicals (·OH), while in the reduction reaction the electrons reacted with O 2 to form superoxide radicals (O 2 − ) [47]. Furthermore, in previous work, we indicated that the BFCTO-1.50 sample showed the highest photocatalysis efficiency, which may have been due to the nanoplates in the BFCTO-1.50 sample having the appropriate area ratio between the {001} facets and the {117} facets. This resulted in a similar reaction ratio on these two facets, thereby efficiently inhibiting the recombination of the charge carrier [23]. Consequently, we predicted that the BFCTO-1.5/NGF would show the best photocatalytic activity, which could be attributed to the effect of {001} and {117} crystal facets with the optimal ratio. which are fully in the BFCTO absorption range, leading to photoexcitation of the BFCTO nanoparticles in BFCTO/NGF. Therefore, the holes (h + ) on the valence band (VB) and the electrons (e − ) on the conduction band (CB) were able to be generated to decompose the RhB molecules. The oxidative holes could react with H2O to produce hydroxyl radicals (•OH), while in the reduction reaction the electrons reacted with O2 to form superoxide radicals (O2 − ) [47]. Furthermore, in previous work, we indicated that the BFCTO-1.50 sample showed the highest photocatalysis efficiency, which may have been due to the nanoplates in the BFCTO-1.50 sample having the appropriate area ratio between the {001} facets and the {117} facets. This resulted in a similar reaction ratio on these two facets, thereby efficiently inhibiting the recombination of the charge carrier [23]. Consequently, we predicted that the BFCTO-1.5/NGF would show the best photocatalytic activity, which could be attributed to the effect of {001} and {117} crystal facets with the optimal ratio. The magnetic properties of all the BFCTO/NGF nanocomposites at room temperature were also measured to investigate their potential as a magnetic recyclable catalyst. The magnetization hysteresis (M-H) loops are shown in Figure 8a. It can be seen that all the samples exhibit spontaneous magnetic moments, indicating their ferromagnetic nature. Among the three samples, BFCTO-2.0/NGF exhibited a maximum remnant magnetism (2Mr) of ~0.32 emu/g, and the coercive The magnetic properties of all the BFCTO/NGF nanocomposites at room temperature were also measured to investigate their potential as a magnetic recyclable catalyst. The magnetization hysteresis (M-H) loops are shown in Figure 8a. It can be seen that all the samples exhibit spontaneous magnetic moments, indicating their ferromagnetic nature. Among the three samples, BFCTO-2.0/NGF exhibited a maximum remnant magnetism (2M r ) of~0.32 emu/g, and the coercive field 2H c of BFCTO-2.0/NGF is~1092 Oe. Due to the paramagnetic properties of the NGF nanoparticles (insert of Figure 8a), the BFCTO/NGF samples showed unsaturated magnetization when a high magnetic field was applied. Additionally, Figure 8b,c displays the photos of BFCTO-1.0/NGF solution without and with a nearby magnet of 0.1 T, which indicates that the BFCTO/NGF samples have the potential as novel, recyclable and efficient NIR light-driven photocatalysts.

Conclusions
In summary, the BFCTO/NGF nanocomposites were successfully synthesized by the hydrothermal method followed by an assembly method, wherein the size and morphology of the BFCTO component was well-adjusted by changing the concentration of NaOH. Under the NIR irradiation of 980nm by a continuous wave diode laser, all the BFCTO samples exhibited no activity. In contrast, with the loading of the NGF upconversion nanoparticles, high photocatalytic activity was observed in the BFCTO/NGF samples, indicating that the loading of NGF nanoparticles could extend the absorption of BFCTO from Vis to NIRlight. Among the three samples, BFCTO-1.5/NGF displayed the best NIR-responsive photocatalysis activity, which could be attributed to the effect of {001} and {117} crystal facets with the optimal ratio. Furthermore, the ferromagnetic property of the BFCTO/NGF sample indicates itspotential as anovel, recyclable and efficient NIR light-driven photocatalyst.
Supplementary Materials: Figure S1: The FT-IR spectra of oleic acid, oleic-capped NGF and oleic-free NGF nanoparticles.

Conclusions
In summary, the BFCTO/NGF nanocomposites were successfully synthesized by the hydrothermal method followed by an assembly method, wherein the size and morphology of the BFCTO component was well-adjusted by changing the concentration of NaOH. Under the NIR irradiation of 980 nm by a continuous wave diode laser, all the BFCTO samples exhibited no activity. In contrast, with the loading of the NGF upconversion nanoparticles, high photocatalytic activity was observed in the BFCTO/NGF samples, indicating that the loading of NGF nanoparticles could extend the absorption of BFCTO from Vis to NIRlight. Among the three samples, BFCTO-1.5/NGF displayed the best NIR-responsive photocatalysis activity, which could be attributed to the effect of {001} and {117} crystal facets with the optimal ratio. Furthermore, the ferromagnetic property of the BFCTO/NGF sample indicates itspotential as anovel, recyclable and efficient NIR light-driven photocatalyst.

Conflicts of Interest:
The authors declare no conflict of interest.