Direct Z-scheme GdFeO3/g-C3N4 heterostructures with outstanding photocatalytic activity

The direct Z-scheme GdFeO 3 /g-C 3 N 4 photocatalysts used without any mediators were obtained using an ultrasonic synthesis approach. The composites’ structures, elemental composition and morphologies, as well as optical, electrochemical and catalytic properties, were thoroughly analyzed. GdFeO 3 /g-C 3 N 4 -3:7 composite exhibited an excellent enhanced light-harvesting ability and low recombination rate, which manifested into an outstanding visible-light driven activity, 7.7 and 2.3 times more than of stand-alone g-C 3 N 4 and GdFeO 3 , respectively. A mechanism responsible for the photocatalytic property enhancement was proposed. The stability and reusability of the GdFeO 3 /g-C 3 N 4 composites make them excellent materials for practical applications related to the removal of organic pollutants from the wastewater. This work reported the preparation of GdFeO 3 /g-C 3 N 4 catalyst with various GdFeO 3 :g-C 3 N 4 ratios. We used a simple ultrasonication-assisted synthesis approach. The resulting hybrid nanocomposites were tested for their photocatalytic activity towards RhB decomposition. GFO/CN-3:7 showed photocatalytic activity signicantly higher than other composites and than unmodied GdFeO 3 and g-C 3 N 4 : 92% of RhB decomposed when the solution was exposed to the visible light for 120 mins. The photocatalytic mechanism was described as a Z-scheme with the •O 2– , •OH, and h + being the most active species. EIS and PL data showed that such excellent activity was because of the very ecient separation of the photogenerated electron-hole pairs, which was accelerated by the interfacial charge carrier recombination. The results obtained in this work provide novel ideas on the development of very ecient visible-light photocatalysts and the usage of GdFeO 3 /g-C 3 N 4 composites for organic pollutant removal.


Introduction
Environment deterioration, including water contamination, is becoming more and more alarming as the energy consumed by the population and its industrial activities increase [1,2]. The photocatalytic assistance of nano-semiconductors attracts a lot of interest in pollutant degradation and removal from water and hydrogen energy production because of their e ciency, cost-effectiveness, and environmental friendliness [3,4]. However, the development of the most common nano-photocatalysts (such as TiO 2 , CdS, and WO 3 ) is hindered by their photo-corrosion, low sensitivity to the visible light, and rapid recombination of photogenerated electron-hole pairs [5,6]. Many research and development efforts are currently undertaken to improve these materials by doping, creating heterojunctions, and surface modi cation [7][8][9].
Perovskite-type ferrite oxides, such as LnFeO 3 (Ln = La, Sm, Y, Dy, Gd, Nd, etc.), are widely used as isolators, gas sensors, and optical switches as well as for enhancing magnetic materials and lithium-ion batteries because of their low density, excellent stability, low toxicity, inertness, and ease of production [10,11]. They also possess bandgaps (in the 1.9-2.5 eV range) suitable for visible-light-driven photocatalysis. Among these compounds is a GdFeO 3 (GFO) perovskite, which contains Gd 3+ with a halffull 4f subshell and possesses stable structure, narrow bandgap, and excellent optical absorption and photocatalytic properties [12][13][14].
Carbon nitride (g-C 3 N 4 ) with a layered graphite-like structure is a π-conjugated semiconductor with a narrow bandgap (equal to ~2.7 eV) and excellent photoactivity under visible light irradiation [15]. Its activity and quantum e ciency can be enhanced even more by doping and combining with other compounds. In fact, g-C 3 N 4 /Ag 2 CrO 4 , g-C 3 N 4 /BiOBr, BiOCl/g-C 3 N 4, g-C 3 N 4 /BiFeO 3 , CeVO 4 /g-C 3 N 4 , TiO 2 /g-C 3 N 4 , and YFeO 3 /g-C 3 N 4 heterostructures were successfully fabricated [16][17][18][19][20][21][22]. This paper reports the fabrication of novel magnetic GdFeO 3 /g-C 3 N 4 heterostructures, which, to the best of our knowledge, were fabricated and used for dye degradation reaction for the rst time. We used a simple ultrasonic synthesis method to obtain heterostructures with a range of GdFeO 3 :g-C 3 N 4 ratios. The heterogeneous visible-light catalytic activity of these heterostructures was tested using Rhodamine B (RhB). We also characterized the ability of these heterostructures to scavenge reactive species as well as their stability and reusability.

Experimental Section
All initial chemicals were analytically pure and used as received. GdFeO 3 perovskite oxide power has been synthesized by the glycol-assisted sol-gel rapid calcination process in the orthorhombic phase with the pbnm space group as reported by our research group [23]. 15 mmol of Gd(NO 3 ) 3 ·6H 2 O and 15 mmol of Fe(NO 3 ) 3 ·9H 2 O were stirred with 40 mL of ethylene glycol for 4 h at 80 °C, after which the sol was dried for 8 hours at 120 °C. The resultant material was calcined at 800 °C for 4 h, after which it was rinsed and dried. The resulting product was brown GdFeO 3 nano-powder. g-C 3 N 4 nanosheets were synthesized using 5 g of melamine heated at 520 °C for 4 h in a mu e furnace. The product was washed for several times with distilled water and absolute ethanol and then at 80 °C for 10 h [24].

Characterization
Crystallinity and structures of the samples were characterized by the X-ray diffraction (XRD) conducted using Bruker-AXS D8 instrument operated with Cu K α -radiation at 40 kV and 40 mA. The spectra were collected in the 20-80 2θ range. GFO/CN-3:7 sample was also characterized by a transmission electron Hitachi Model H-800 microscope (TEM) operated at 200 kV. X-ray photoelectron spectroscopy (XPS) was conducted by a Perkin-Elmer PHI-1600 ESCA instrument using an Mg Ka radiation as an X-ray source and a C 1s peak energy (equal to 284.6 eV) for calibration. The Purkinje General T1901 UV-Vis spectrophotometer (containing an integrating sphere attachment and BaSO 4 re ectance standard) was used to collect adsorption spectra. The photoluminescence (PL) data were obtained by the uorescence F4600 spectrometer manufactured by Hitachi (Japan) using a 300 nm excitation wavelength. The electrochemical impedance spectra (EIS) were collected by a CHI660E workstation using a three-electrode system and 0.5 M KCl electrolyte.

Photocatalytic tests
The catalytic activities of our materials were evaluated at 25 ºC using rhodamine B (RhB) degradation reaction by a Zhengqiao ZQ-GHX-V apparatus. 500 W Xe lamp with an optical cutoff lter below 420 nm was used for sample irradiation. For this purpose, 10 mg of a sample was mixed with the 10 mL of aqueous RhB solution and stirred in the dark for 1 h. Then, the reaction vessel was exposed to the visible light. Aliquots were taken out and centrifuged. The supernatant was then subjected to the UV-Vis spectrophotometry (performed using T1901 instrument manufactured by Purkinje General) to determine the amount of RhB (which was performed at its characteristic wavelength of 553 nm). The RhB degradation e ciency C/C 0 was calculated using initial (C 0 ) and time-dependent (C) RhB concentrations.
The UV-vis diffuse re ectance spectroscopy (DRS) was used to assess light absorption. TEM of the GFO/CN-3:7 showed GdFeO 3 nanoparticles uniformly distributed on the g-C 3 N 4 nanosheets (see Fig. 2(a) and (b)). HRTEM revealed lattice fringes 0.273 nm apart (which correspond to the (112) plane of orthorhombic GdFeO 3 ) and very distinct g-C 3 N 4 /GdFeO 3 interface (see Fig. 2(c)). Thus, g-C 3 N 4 nanosheets served as substrates for GdFeO 3 nanoparticle growth and also prevented their agglomeration and growth into larger particles. The presence of numerous nanoparticles is bene cial for a catalyst because they provide a large surface area. Additionally, a tight g-C 3 N 4 /GdFeO 3 interface is advantageous for the fast and e cient charge transfers [25].

Results And Discussion
XPS of the GFO/CN-3:7 sample showed two peaks centered at 710.6 and 724.4 eV, ascribed to Fe 2p 3/2 and Fe 2p 1/2 , respectively (see Fig 3a). Thus, Fe(III) was in its oxide coordination [26]. Gd core level spectrum showed 4d peaks located at 141.6 and 147.3 eV (see Fig. 3b). The main feature of the Gd 4d photoemission arose from multiplet splitting of the 4d hole with 4f 7 valence electrons, forming 9 D and 7 D nal ionic states [27,28]. The main O form is O 2-, as indicated by the O 1s spectrum, consisting of two peaks at 529.2 and 531.3 eV (see Fig. 3c). These peaks correspond to the hydroxyl species (O H ) and the GdFeO 3 crystal lattice oxygen (O L , both in Gd-O and Fe-O con gurations) [29]. The broad and asymmetric N 1s peak was deconvoluted into peaks positioned at 399.64, 399.03, and 398.14 eV (see Fig. 3d), which were interpreted as belonging to the H-N-(C) 2 and N-(C) 3 as well as to C-N=C bonds of g-C 3 N 4 [30,31], respectively. The N 1s peak observed for the GdFeO 3 /g-C 3 N 4 was slightly shifted relative to the g-C 3 N 4 peak position at 398.20 eV [32,33]. Thus, the composite material possessed a somewhat different chemical environment than stand-alone g-C 3 N 4 because of the GdFeO 3 and g-C 3 N 4 interaction. GdFeO 3 and g-C 3 N 4 are capable of absorbing light up to 685 and 455 nm, respectively (see Fig. 4a). Thus, GdFeO 3 absorbs both visible and UV light. The GdFeO 3 /g-C 3 N 4 absorption intensity decreased as the g-C 3 N 4 content increased. Additionally, the absorption edge of the GdFeO 3 /g-C 3 N 4 composites red-shifted relative to unmodi ed GdFeO 3 as the GdFeO 3 content in the sample increased. The wide light absorption range of the GdFeO 3 /g-C 3 N 4 composites was because of its bandgap decrease (relative to unmodi ed g-C 3 N 4 ) from 2.83 to 2.22-2.02 eV (see Fig. 4b) [34].
The maximum of the UV-Vis absorption RhB under the presence of GFO/CN-3:7 was at 553 nm (see Fig.  5a). The peak intensity decreased with time, which con rms RhB decomposition under the GFO/CN-3:7 presence (also corroborated by the gradual RhB solution discoloration visible by the naked eye). A blueshifted absorption peak was ascribed to the RhB deethylation [35]. Photodegradation e ciency of unmodi ed g-C 3 N 4 and GdFeO 3 was low: only 9.5% and 41% of RhB was removed after 120 min of visible-light irradiation (see Fig. 5b). At the same time, GFO/CN-3:7 composite decomposed 92% of RhB (see Fig. 5b), which was the highest value among all other composites obtained in this work. The conclusion of experiment proves that exists an optimum composite ratio between GdFeO 3 and g-C 3 N 4 , which yields best e ciency [36].
We also performed photocatalytic reactions of RhB decomposition assisted by the GFO/CN-3:7 under the presence of radical-and hole-capturing compounds to understand which active oxygen species (AOSs) were dominant (see Fig. 6a). For this purpose, we added isopropyl alcohol (IPA) and p-benzoquinone (BQ) to quench hydroxyl (•OH) and superoxide (•O 2 − ) radicals, respectively [37]. Only 5.5% and almost 0% of RhB was removed after 120 min under the IPA and BQ presence (see Fig. 6). Thus, the main AOSs were •OH and •O 2 − . We also added disodium ethylenediaminetetraacetate (EDTA-2Na) to quench holes and observed the decomposition of only 25.7% of RhB [37]. which means that the dye molecules can also be oxidated directly by the photogenerated holes occurs from separation of photoproduction electron-hole pairs. spectra of GdFeO 3 showed almost no emission, while g-C 3 N 4 showed strong uorescence (see Fig. 7(a)).
PL intensity of the GFO/CN composites was lower than that of g-C 3 N 4 . Thus, the electron-hole pair separation was because of g-C 3 N 4 coupling with GdFeO 3 . Out of all composites, GFO/CN-3:7 demonstrated the lowest PL emission intensity, which is indicative of the highest charge transfer e ciency and the possibility of the highest photocatalytic oxidation activity [38]. Fig. 7(b) shows EIS data obtained for g-C 3 N 4 , GdFeO 3, and GFO/CN-3:7. The radii of the semicircles are proportional to the charge transfer resistance. GFO/CN-3:7 exhibited the smallest semicircle, which is indicative of the best charge transfer e ciency and conductivity out of these three samples, which agrees with the PL data [39,40].
To be feasible for practical applications, a catalyst must be stable and reusable. To demonstrate this, we used GFO/CN-3:7 composite for three consecutive photocatalytic RhB degradation reactions performed under the same conditions (see Fig. 8). During the rst, second and third reactions, GFO/CN-3:7 was able to assist in decomposition of 92, 91 and 87% of the initial RhB. A slight decline of the RhB removal rate by the GFO/CN-3:7 catalyst was attributed to not fully recovered active sites, and minor weight loss occurred during post-reaction rinsing.
To explain such excellent photocatalytic activity of the GFO/CN-3:7 nanocomposite, we rst used the concept of the band-edge potential levels and heterojunctions, which determines the separation and recombination of photoelectron-hole pairs. GdFeO 3 valence and conduction band (VB and CB, respectively) potentials were lower than those of g-C 3 N 4 . Thus, the electrons can travel from the g-C 3 N 4 CB to the GdFeO 3 CB. Simultaneously, the photogenerated holes can move from the GdFeO 3 VB to the g-C 3 N 4 VB. However, the E CB of GdFeO 3 is above the O 2 /•O 2 − potential (which is equal to -0.33 eV vs. NHE).
Therefore, the hole transport from the GdFeO 3 VB is very slow, which lowers that material's photocatalytic activity.
Therefore, we considered a Z-scheme mechanism instead. The photogenerated electrons on the GdFeO 3 CB transfer with ease to the g-C 3 N 4 VB, reacting with the g-C 3 N 4 holes and inhibiting their recombination in both GdFeO 3 and g-C 3 N 4 . This, in turn, prolongs the time electrons and holes remain on the g-C 3 N 4 CB and GdFeO 3 VB, respectively. These processes enhance the charge transfer at the g-C 3 N 4 /GdFeO 3 interface (see Fig. 9). At the same time, the electrons from g-C 3 N 4 CB with strong reduction potential (equal to -1.14 eV react with O 2 producing •O 2 − (see Eq. (1)).

Conclusions
This work reported the preparation of GdFeO 3 /g-C 3 N 4 catalyst with various GdFeO 3 :g-C 3 N 4 ratios. We used a simple ultrasonication-assisted synthesis approach. The resulting hybrid nanocomposites were tested for their photocatalytic activity towards RhB decomposition. GFO/CN-3:7 showed photocatalytic activity signi cantly higher than other composites and than unmodi ed GdFeO 3 and g-C 3 N 4 : 92% of RhB decomposed when the solution was exposed to the visible light for 120 mins. The photocatalytic mechanism was described as a Z-scheme with the •O 2 -, •OH, and h + being the most active species. EIS and PL data showed that such excellent activity was because of the very e cient separation of the photogenerated electron-hole pairs, which was accelerated by the interfacial charge carrier recombination. The results obtained in this work provide novel ideas on the development of very e cient visible-light photocatalysts and the usage of GdFeO 3 /g-C 3 N 4 composites for organic pollutant removal. Declarations