Improved Visible-Light Photocatalytic H2 Evolution of G-C3N4 Nanosheets by Constructing Heterojunctions with Nano-Sized Poly(3-Thiophenecarboxylic Acid) and Coordinating Fe(III)

It is highly desirable to enhance the photogenerated charge separation of g-C3N4 by constructing efficient heterojunctions, especially with an additional organic constitution for solar–hydrogen conversion. Herein, g-C3N4 nanosheets have been modified controllably with nano-sized poly(3-thiophenecarboxylic acid) (PTA) through in situ photopolymerization and then coordinated with Fe(III) via the -COOH groups of modified PTA, forming an interface of tightly contacted nanoheterojunctions between the Fe(III)-coordinated PTA and g-C3N4. The resulting ratio-optimized nanoheterojunction displays a ~4.6-fold enhancement of the visible-light photocatalytic H2 evolution activity compared to bare g-C3N4. Based on the surface photovoltage spectra, measurements of the amount of •OH produced, photoluminescence (PL) spectra, photoelectrochemical curves, and single-wavelength photocurrent action spectra, it was confirmed that the improved photoactivity of g-C3N4 is attributed to the significantly promoted charge separation by the transfer of high-energy electrons from the lowest unoccupied molecular orbital (LUMO) of g-C3N4 to the modified PTA via the formed tight interface, dependent on the hydrogen bond interaction between the -COOH of PTA and the -NH2 of g-C3N4, and the continuous transfer to the coordinated Fe(III) with -OH favorable for connection with Pt as the cocatalyst. This study demonstrates a feasible strategy for solar-light-driven energy production over the large family of g-C3N4 heterojunction photocatalysts with exceptional visible-light activities.


Introduction
Hydrogen is a potential energy carrier that is high energy density and carbon-free, making it a splendid candidate for energy storage and conversion. Photocatalytic water splitting is one of the simplest pathways to mimic photosynthesis for the conversion of solar energy into storable, clean and renewable H 2 fuels, which can well relieve the pressure caused by environmental and energy problems [1,2]. It is well-known that in the clean and environmentally friendly photocatalytic process, a semiconductor photocatalyst is excited by light to generate electrons and holes so as to induce reduction and oxidation reactions. Thus, the semiconductor photocatalyst is the key factor affecting the photocatalytic reaction [3]. Since visible light comprises large proportion (approximately 44%) of the sunlight that reaches the Earth's surface, from the perspective of practical applications, it is of great significance to develop photocatalysts that can respond in the wide visible range. For this reason researchers have conducted significant research and exploration in the field of photocatalysts with a visible-light response [4][5][6].
with H 2 O molecules and thus enhancing the photocatalytic H 2 O reduction performance for H 2 evolution [30,31]. Inspired by these factors, it is highly anticipated that PTA will be introduced to construct tightly contacted heterojunctions with g-C 3 N 4 for promoting the photogenerated charge transfer and separation, extending the visible-light response, improving the water dispersibility and thus enhancing the visible-light photocatalytic H 2 evolution activities.
For photocatalysts, the coordinated transition metal ions play an important role in their photogenerated charge transfer and separation. Sardar et al. revealed an efficient photoinduced electron-transfer processes in Fe(III)-hematoporphyrin-TiO 2 nanohybrids from excited hematoporphyrin to Fe(III), which was responsible for the improved photocatalytic performance [32]. Zhang et al. proved that the photogenerated electrons of g-C 3 N 4 transferred to tetra(4-carboxyphenyl)porphyrin iron(III) chloride (FeTCPP), leading to the formation of Fe II TCPP, [Fe II TCPP . . . CO 2 ], [Fe I TCPP . . . CO 2 ] − and [Fe 0 TCPP . . . CO 2 ] 2− intermediates in g-C 3 N 4 /FeTCPP heterogeneous catalysts upon irradiation, thus the improved photocatalytic activity for CO 2 reduction [28]. As one of the interesting coordinating moieties, -COOH is widely employed for the construction of transition metal carboxylate clusters by coordinating with transition metal ions, using either one or both oxygen atoms [33,34]. Based on the above considerations, it is highly feasible to construct Fe(III)-coordinated PTA/g-C 3 N 4 heterojunctions by coordinating Fe(III) with the -COOH of PTA. This could improve the photogenerated charge transfer and separation and thus the visible-light photocatalytic activities. As far as we know, Fe(III)-coordinated PTA/g-C 3 N 4 heterojunctions for visible-light photocatalytic H 2 evolution have not been reported.
In this work, PTA/g-C 3 N 4 nanoheterojunctions with tight interface contact were first successfully synthesized using the in situ photopolymerization method. Fe(III)-coordinated PTA/g-C 3 N 4 nanoheterojunctions were then constructed using the simple impregnation method with an Fe(NO 3 ) 3 solution. It was proven that the tight interface contact, dependent on the hydrogen bond interaction between the -COOH of PTA and the -NH 2 of g-C 3 N 4 , is responsible for the promoted photogenerated charge separation and thus the enhanced visible-light catalytic H 2 evolution activities and the visible photosensitization of the PTA. Interestingly, by coordinating with the -COOH of PTA, the introduced Fe(III) ions further improve the photogenerated charge separation and visible-light catalytic H 2 evolution activities. By rational optimization, the resulting 0.5Fe-2PTA/g-C 3 N 4 shows H 2 evolution activity of up to 687.3 µmol h −1 g −1 under irradiation with visible light, which is approximately a~4.6-fold enhancement of g-C 3 N 4 . Moreover, single-wavelength photocurrent action spectra illustrate that the promoted charge separation is mainly attributed to the high-level electron transfer from the LUMO of g-C 3 N 4 to the LUMO of PTA. This work offers a practicable route to design and fabricate visible-light-driven g-C 3 N 4 heterojunction photocatalysts for solar-hydrogen conversion.

Synthesis of G-C 3 N 4 Nanosheets
G-C 3 N 4 nanosheets were synthesized according to our previous report [35]. Typically, a certain amount of urea was placed into an alumina combustion boat with a cover, and was then heated to 550 • C for 3 h at a rate of 0.5 • C min −1 in a muffle furnace in an air atmosphere. After being cooled to room temperature naturally, the light-yellow product was collected and ground into powder.

Synthesis of PTA/g-C 3 N 4 Nanoheterojunctions
PTA/g-C 3 N 4 nanoheterojunctions were synthesized by the in situ photopolymerization method, according to the literature [36,37]. A certain amount of g-C 3 N 4 and 3-thiophenecarboxylic acid (TA) was dispersed in 150 mL deionized water, stirred for 1 h and then treated by ultrasonication for 30 min. The potassium dichromate (K 2 Cr 2 O 7 ), with a 1:8 molar ratio of TA to K 2 Cr 2 O 7 , was then added into the above suspension with stirring for another 1 h. The mixture was irradiated under visible light (λ > 420 nm) for  48 h in a cooled water bath. The PTA/g-C 3 N 4 nanoheterojunctions, obtained by filtering, were washed with 1 M HCl and deionized water three times, dried overnight at 60 • C in a vacuum and ground into powder for further use. The obtained nanoheterojunctions were denoted as xPTA/g-C 3 N 4 , where x represents the mass percent of PTA in the PTA/g-C 3 N 4 nanoheterojunctions (x = 1, 2, 3 and 5 wt%). For comparison, the PTA was prepared by the photopolymerization without g-C 3 N 4 under the same other conditions.

Synthesis of Fe(III)-Coordinated PTA/g-C 3 N 4 Nanoheterojunctions
Fe(III)-coordinated PTA/g-C 3 N 4 nanoheterojunctions were fabricated using a simple impregnation method. In a typical procedure, a certain amount of 2PTA/g-C 3 N 4 was dispersed in 100 mL aqueous solution with a certain amount of Fe(NO 3 ) 3 ·9H 2 O as the iron source, stirred for 48 h, filtered and then dried at 60 • C overnight in a vacuum. The samples were ground into powder and denoted as yFe-2PTA/g-C 3 N 4 for further use, where y represents the mass percent of the Fe(III) in the yFe-2PTA/g-C 3 N 4 nanoheterojunctions (y = 0.1, 0.3, 0.5, 0.8 and 1.0 wt%). For comparison, 0.5Fe-g-C 3 N 4 and 0.5Fe-PTA with 0.5 wt% Fe(III) in the samples were prepared through the same impregnation method without PTA or g-C 3 N 4 under the same other conditions, respectively.
For comparison, Fe(III)-polythiophene/g-C 3 N 4 heterojunctions were fabricated in which there was no -COOH on the main chain of polythiophene (PTh), so there was no hydrogen bond or coordination bond interaction between the PTh and g-C 3 N 4 or Fe(III). Firstly, 2PTh/g-C 3 N 4 was synthesized through the in situ chemical oxidation method. Typically, 980 mg g-C 3 N 4 was dispersed in 150 mL acetonitrile solution, stirred for 1 h and then treated by ultrasonication for 30 min. A total of 19.1 µL thiophene was added into the above suspension with stirring for another 1 h, and then 5 mL acetonitrile solution containing 77.1 mg ferric chloride was added into the suspension dropwise and stirred in an ice bath for 3 h. The 2PTh/g-C 3 N 4 , obtained by filtering, was washed with anhydrous methanol and deionized water three times and then dried overnight at 60 • C in a vacuum. Secondly, 0.5Fe-2PTh/g-C 3 N 4 was prepared using the same impregnation method. In detail, 300 mg 2PTh/g-C 3 N 4 was dispersed in 100 mL aqueous solution containing 10.8 mg Fe(NO 3 ) 3 ·9H 2 O, stirred for 48 h, filtered and then dried overnight at 60 • C in a vacuum. The obtained sample was ground into powder and denoted as 0.5Fe-2PTh/g-C 3 N 4 .

Evaluation of Photocatalytic Activities for H 2 Evolution
The H 2 evolution measurements were carried out in a 250 mL quartz cell with a 300 W Xe lamp with a cut-off filter (λ > 420 nm). Typically, 100 mg sample was added into the reactor, in addition to 100 mL of 10 vol% triethanolamine (TEOA) aqueous solution, which acted as the cavitation scavenger. Before irradiation, the system was vacuumed to remove dissolved air. Then, 1 wt% Pt was loaded onto the surface of the photocatalyst as a cocatalyst by the in situ photodeposition of H 2 PtCl 6 ·6H 2 O for 2 h. The sample was then irradiated in a closed water-circulating system for 5 h. The amount of evolved H 2 was detected with an online TCD gas chromatograph (GC-7900, Techcomp, Shanghai, China), using nitrogen as the carrier gas. The stability of the samples was measured for 20 h with a 5 h run cycle.
The apparent quantum yield (AQY) for the evolution of H 2 was calculated as follows: where M is the amount of H 2 produced in the reaction (mol), N A is Avogadro's constant (6.02 × 10 23 /mol), h is the Planck constant (6.626 × 10 −34 J S), c is the vacuum light velocity (3 × 10 8 m/s), S is the irradiation area (cm 2 ), P is the monochromatic light intensity (W/cm 2 ), t is the photoreaction time (s) and λ is the wavelength of the monochromatic light (m).
A characterization of the materials, photoelectrochemical and electrochemical measurements, and an evaluation of the amount of •OH samples produced are described in the supporting information.

Results and Discussion
The fabricating strategy and synthetic route for the Fe(III)-coordinated PTA/g-C 3 N 4 nanoheterojunctions are shown in Figure 1 and Scheme S1. G-C 3 N 4 nanosheets were first prepared by the thermal polymerization of urea. Subsequently, TA was linked to the g-C 3 N 4 nanosheets by a hydrogen bond interaction between the -COOH of PTA and the -NH 2 of g-C 3 N 4 . The tightly contacted PTA/g-C 3 N 4 nanoheterojunctions were then constructed by the in situ photopolymerization of TA under irradiation with visible light. Finally, Fe(III) was introduced by coordinating with the -COOH of PTA to fabricate Fe(III)-coordinated PTA/g-C 3 N 4 nanoheterojunctions, which are expected to be helpful for improving photogenerated charge separation and visible-light photocatalytic H 2 evolution activities.
where M is the amount of H2 produced in the reaction (mol), NA is Avogadro's constant (6.02 × 10 23 /mol), h is the Planck constant (6.626 × 10 −34 J S), c is the vacuum light velocity (3 × 10 8 m/s), S is the irradiation area (cm 2 ), P is the monochromatic light intensity (W/cm 2 ), t is the photoreaction time (s) and λ is the wavelength of the monochromatic light (m).
A characterization of the materials, photoelectrochemical and electrochemical measurements, and an evaluation of the amount of •OH samples produced are described in the supporting information.

Results and Discussion
The fabricating strategy and synthetic route for the Fe(III)-coordinated PTA/g-C3N4 nanoheterojunctions are shown in Figure 1 and Scheme S1. G-C3N4 nanosheets were first prepared by the thermal polymerization of urea. Subsequently, TA was linked to the g-C3N4 nanosheets by a hydrogen bond interaction between the -COOH of PTA and the -NH2 of g-C3N4. The tightly contacted PTA/g-C3N4 nanoheterojunctions were then constructed by the in situ photopolymerization of TA under irradiation with visible light. Finally, Fe(III) was introduced by coordinating with the -COOH of PTA to fabricate Fe(III)coordinated PTA/g-C3N4 nanoheterojunctions, which are expected to be helpful for improving photogenerated charge separation and visible-light photocatalytic H2 evolution activities.

Structure Characterization
As presented in Figure S1a, it is clear that g-C3N4 displays two characteristic diffraction peaks at around 13.0° and 27.4°, corresponding to its (001) and (002) facets (JCPDS 87-1526) [38]. Additionally, PTA shows a wide peak at 24.9° ( Figure S1b), which is attributed to the π-π stacking of the conjugate polymerization backbone [39]. By comparison, PTA/g-C3N4 and 0.5Fe-2PTA/g-C3N4 display similar diffraction peaks, indicating that the lattice structure of g-C3N4 does not change after coupling with PTA and coordinating with Fe(III) [21]. Notably, no apparent diffraction peaks of PTA are observed in any nanoheterojunctions, which can be ascribed to the low crystallinity and content of PTA [40]. The UV-Vis diffuse reflection (UV-Vis DRS, Shimadzu, Kyoto, Japan) spectra ( Figure 2a) show that the light absorption range of 2PTA/g-C3N4 is expanded compared to g-C3N4, and there is no further change after coordinating with Fe(III), indicating that the introduc-

Structure Characterization
As presented in Figure S1a, it is clear that g-C 3 N 4 displays two characteristic diffraction peaks at around 13.0 • and 27.4 • , corresponding to its (001) and (002) facets (JCPDS 87-1526) [38]. Additionally, PTA shows a wide peak at 24.9 • (Figure S1b), which is attributed to the π-π stacking of the conjugate polymerization backbone [39]. By comparison, PTA/g-C 3 N 4 and 0.5Fe-2PTA/g-C 3 N 4 display similar diffraction peaks, indicating that the lattice structure of g-C 3 N 4 does not change after coupling with PTA and coordinating with Fe(III) [21]. Notably, no apparent diffraction peaks of PTA are observed in any nanoheterojunctions, which can be ascribed to the low crystallinity and content of PTA [40]. The UV-Vis diffuse reflection (UV-Vis DRS, Shimadzu, Kyoto, Japan) spectra ( Figure 2a) show that the light absorption range of 2PTA/g-C 3 N 4 is expanded compared to g-C 3 N 4 , and there is no further change after coordinating with Fe(III), indicating that the introduction of Fe(III) does not affect the optical absorption. As shown in Figure S2, with the increase in the PTA content, the light absorption intensity of the xPTA/g-C 3 N 4 in the visible light range clearly shows a gradual enhancement. It fully demonstrates that the visible-light response of g-C 3 N 4 is significantly expanded due to the visible photosensitization effect of the introduced PTA. tion of Fe(III) does not affect the optical absorption. As shown in Figure S2, with the increase in the PTA content, the light absorption intensity of the xPTA/g-C3N4 in the visible light range clearly shows a gradual enhancement. It fully demonstrates that the visiblelight response of g-C3N4 is significantly expanded due to the visible photosensitization effect of the introduced PTA. To investigate the interface interactions between PTA and g-C3N4 in PTA/g-C3N4, Fourier-transform infrared (FTIR, Bruker, Karlsruhe, Germany) spectra of the samples were analyzed, as shown in Figures 2b and S3. As shown in Figure S3a, PTA shows the carboxylic O-H and C=O stretching peaks at 3400 and 1610 cm −1 , C-H stretching peaks at 3098, 2965 and 2915 cm −1 , thiophene ring vibration peaks at 1533, 1437 and 1356 cm −1 and the C-S stretching vibration peak of the thiophene ring at 902 cm −1 [41][42][43][44]. It is worth noting that PTA presents a substantial absorption peak at 788 cm −1 , corresponding to Cβ-H out-of-plane vibration, indicating that the α-α coupling is the dominant form of C-C bonding for the as-prepared PTA and there is almost no α-β coupling form [45]. As illustrated in Figure 2b, g-C3N4 shows a broad peak at around 3000-3550 cm −1 , originating from the terminal -NH2. The peak at around 810 cm −1 is assigned to the hepazine ring bending vibration, and the set of peaks from 1236 cm −1 to 1650 cm −1 is ascribed to the C-N heterocycle stretching vibration [46,47]. By comparison, in xPTA/g-C3N4 (Figures 2b and S3), the -NH2 absorption peak moves to a lower direction, indicating the interaction between the -COOH of PTA and the -NH2 of g-C3N4. Compared with 2PTA/g-C3N4, 0.5Fe-2PTA/g-C3N4 with coordinated Fe(III) presents an -NH2 absorption peak with no significant change in position, indicating that there may be no coordination bond interaction between the Fe(III) and -NH2 or that the interaction is negligible, which is consistent with another report [48]. Note that in Figure 2b, compared with that of g-C3N4, the peak intensity of 2PTA/g-C3N4 (at 3000-3550 cm −1 ) is strengthened due to the introduction of PTA with -COOH, and the peak intensity is further strengthened in 0.5Fe-2PTA/g-C3N4, attributed to the newly introduced Fe(III) with -OH [49].
The morphology of the samples was studied using a scanning electron microscope (SEM, Hitachi, Tokyo, Japan). Figure S4a show that g-C3N4 displays the laminar structure. Figure S4b-e show that for xPTA/g-C3N4, PTA nanoparticles are dispersed on g-C3N4. It can be further observed from the transmission electron microscope (TEM) images that g-C3N4 ( Figure 3a) displays 2D nanosheets and in 0.5Fe-2PTA/g-C3N4, nano-sized PTA particles with a diameter of 60~80 nm are evenly attached to the nanosheet surface of g-C3N4. Additionally, no Fe aggregates can be observed on the g-C3N4 surface, indicating that Fe does not exist as particles or clusters ( Figure 3b). Furthermore, the high-angle annular To investigate the interface interactions between PTA and g-C 3 N 4 in PTA/g-C 3 N 4 , Fourier-transform infrared (FTIR, Bruker, Karlsruhe, Germany) spectra of the samples were analyzed, as shown in Figures 2b and S3. As shown in Figure S3a, PTA shows the carboxylic O-H and C=O stretching peaks at 3400 and 1610 cm −1 , C-H stretching peaks at 3098, 2965 and 2915 cm −1 , thiophene ring vibration peaks at 1533, 1437 and 1356 cm −1 and the C-S stretching vibration peak of the thiophene ring at 902 cm −1 [41][42][43][44]. It is worth noting that PTA presents a substantial absorption peak at 788 cm −1 , corresponding to C β -H out-of-plane vibration, indicating that the α-α coupling is the dominant form of C-C bonding for the as-prepared PTA and there is almost no α-β coupling form [45]. As illustrated in Figure 2b, g-C 3 N 4 shows a broad peak at around 3000-3550 cm −1 , originating from the terminal -NH 2 . The peak at around 810 cm −1 is assigned to the hepazine ring bending vibration, and the set of peaks from 1236 cm −1 to 1650 cm −1 is ascribed to the C-N heterocycle stretching vibration [46,47]. By comparison, in xPTA/g-C 3 N 4 (Figures 2b and S3), the -NH 2 absorption peak moves to a lower direction, indicating the interaction between the -COOH of PTA and the -NH 2 of g-C 3 N 4 . Compared with 2PTA/g-C 3 N 4 , 0.5Fe-2PTA/g-C 3 N 4 with coordinated Fe(III) presents an -NH 2 absorption peak with no significant change in position, indicating that there may be no coordination bond interaction between the Fe(III) and -NH 2 or that the interaction is negligible, which is consistent with another report [48]. Note that in Figure 2b, compared with that of g-C 3 N 4 , the peak intensity of 2PTA/g-C 3 N 4 (at 3000-3550 cm −1 ) is strengthened due to the introduction of PTA with -COOH, and the peak intensity is further strengthened in 0.5Fe-2PTA/g-C 3 N 4 , attributed to the newly introduced Fe(III) with -OH [49].
The morphology of the samples was studied using a scanning electron microscope (SEM, Hitachi, Tokyo, Japan). Figure S4a show that g-C 3 N 4 displays the laminar structure. Figure S4b-e show that for xPTA/g-C 3 N 4 , PTA nanoparticles are dispersed on g-C 3 N 4 . It can be further observed from the transmission electron microscope (TEM) images that g-C 3 N 4 (Figure 3a) displays 2D nanosheets and in 0.5Fe-2PTA/g-C 3 N 4 , nano-sized PTA particles with a diameter of 60~80 nm are evenly attached to the nanosheet surface of g-C 3 N 4 . Additionally, no Fe aggregates can be observed on the g-C 3 N 4 surface, indicating that Fe does not exist as particles or clusters (Figure 3b). Furthermore, the high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) image and the corresponding EDS mappings of 0.5Fe-2PTA/g-C 3 N 4 (Figure 3c) clearly show the distribution of C, N, O, S and Fe elements, verifying the successful synthesis of 0.5Fe-2PTA/g-C 3 N 4 . It is also noted that the Fe and S elements exhibit a similar distribution, confirming the coordination of Fe(III) on PTA.
dark-field scanning transmission electron microscopy (HAADF-STEM) image and the corresponding EDS mappings of 0.5Fe-2PTA/g-C3N4 (Figure 3c) clearly show the distribution of C, N, O, S and Fe elements, verifying the successful synthesis of 0.5Fe-2PTA/g-C3N4. It is also noted that the Fe and S elements exhibit a similar distribution, confirming the coordination of Fe(III) on PTA. In order to verify the interaction between the -COOH of PTA and the -NH2 of g-C3N4 or Fe(III), an X-ray photoelectron spectroscopy (XPS, Shimadzu, Kyoto, Japan) analysis was performed. As shown in Figure S5a, g-C3N4 is mainly composed of C and N, with a small amount of O coming from the adsorbed H2O, and PTA is mainly composed of C, N, O and S elements. Compared with 2PTA/g-C3N4, Fe was also detected in 0.5Fe-2PTA/g-C3N4 in addition to C, N, O and S, which is in good agreement with the result of the EDS mappings. The high-resolution XPS spectra were displayed to examine the chemical state of various elements. The C 1s spectra of g-C3N4, 2PTA/g-C3N4 and 0.5Fe-2PTA/g-C3N4 (Figure 4a) can be fitted into three peaks at 284.6 eV, 286.2 eV and 288.1 eV, attributed to C=C bonds in aromatic structures, the C-OH bond arising from some surface contamination and sp 2 carbon atoms combined with N (C-N3) in the aromatic structure [50]. N 1s spectra of g-C3N4 (Figure 4b) can be fitted with four peaks at 398.8 eV, 400.1 eV, 401.3 eV and 404.1 eV, attributed to the sp 2 -hybridized aromatic N (C-N=C), the tertiary N bonded to C atoms in the form of N-(C)3, C-N-H side groups and π excitation, respectively [51,52]. It is noted that compared with that of g-C3N4, the C-N-H peaks of 2PTA/g-C3N4 and 0.5Fe-2PTA/g-C3N4 shift to the same higher binding energy (401.5 eV), indicating the decreased electron density around the N atoms of g-C3N4 in 2PTA/g-C3N4 and 0.5Fe-2PTA/g-C3N4 due to the In order to verify the interaction between the -COOH of PTA and the -NH 2 of g-C 3 N 4 or Fe(III), an X-ray photoelectron spectroscopy (XPS, Shimadzu, Kyoto, Japan) analysis was performed. As shown in Figure S5a, g-C 3 N 4 is mainly composed of C and N, with a small amount of O coming from the adsorbed H 2 O, and PTA is mainly composed of C, N, O and S elements. Compared with 2PTA/g-C 3 N 4 , Fe was also detected in 0.5Fe-2PTA/g-C 3 N 4 in addition to C, N, O and S, which is in good agreement with the result of the EDS mappings. The high-resolution XPS spectra were displayed to examine the chemical state of various elements. The C 1s spectra of g-C 3 N 4 , 2PTA/g-C 3 N 4 and 0.5Fe-2PTA/g-C 3 N 4 (Figure 4a) can be fitted into three peaks at 284.6 eV, 286.2 eV and 288.1 eV, attributed to C=C bonds in aromatic structures, the C-OH bond arising from some surface contamination and sp 2 carbon atoms combined with N (C-N 3 ) in the aromatic structure [50]. N 1s spectra of g-C 3 N 4 (Figure 4b) can be fitted with four peaks at 398.8 eV, 400.1 eV, 401.3 eV and 404.1 eV, attributed to the sp 2 -hybridized aromatic N (C-N=C), the tertiary N bonded to C atoms in the form of N-(C) 3 , C-N-H side groups and π excitation, respectively [51,52]. It is noted that compared with that of g-C 3 N 4 , the C-N-H peaks of 2PTA/g-C 3 N 4 and 0.5Fe-2PTA/g-C 3 N 4 shift to the same higher binding energy (401.5 eV), indicating the decreased electron density around the N atoms of g-C 3 N 4 in 2PTA/g-C 3 N 4 and 0.5Fe-2PTA/g-C 3 N 4 due to the hydrogen bond interaction between -NH 2 of g-C 3 N 4 and the -COOH of PTA and no obvious interaction between the N atom of g-C 3 N 4 and the Fe(III) in 0.5Fe-2PTA/g-C 3 N 4 [19,21]. As shown in Figure 4c, compared with that of PTA, the O 1s peak of 2PTA/gmoves to a lower binding energy, indicating the increased electron density around t atoms of PTA and confirming the hydrogen bond interaction between the -NH2 of gand the -COOH of PTA, combined with the above result of the C-N-H peak shift (Fi 4b). Interestingly, compared with that of the 2PTA/g-C3N4, the O 1s peak of 0.5Fe-2PT C3N4 exhibits a moderate shift to the higher binding energy direction, indicating the creased electron density around the O atoms of PTA due to the coordination bond i action between the introduced Fe(III) and the -COOH of PTA. The interaction between Fe(III) and the -COOH of PTA was also verified by the O 1s peaks of the PTA and 0 PTA in Figure S5b, which clearly shows the O 1s peak shift of the PTA from 532.1 e 532.4 eV after introducing the Fe(III) [53]. It is worth noting that in Figures 4c and S5b O 1s peaks of the 0.5Fe-2PTA/g-C3N4 and 0.5Fe-PTA are higher than those of the 2PT C3N4 and PTA, respectively, which is due to the -OH of the introduced Fe(III). The F spectra of 0.5Fe-2PTA/g-C3N4 (Figure 4d) can be fitted with Fe 2p3/2 peaks at 710.5 eV 713.4 eV and the corresponding symmetric peaks of Fe 2p1/2 at 723.6 eV and 726.4 eV spectively, while the peaks at 718.7 eV and 731.7 eV are ascribed to the Fe(III) sate peak [35,48]. These are characteristic of Fe 3+ , further confirming the existence of Fe(II 0.5Fe-2PTA/g-C3N4. To further explore the change of the Fe 2p peaks caused by the i action between Fe(III) and the -COOH of PTA, the Fe 2p spectra of 0.5Fe-2PTA, 0.5F C3N4 and 0.5Fe-2PTA/g-C3N4 were also analyzed. As shown in Figure S5c, compared those of 0.5Fe-g-C3N4, the Fe 2p peaks of 0.5Fe-PTA move to a smaller direction, dem strating that the interaction of Fe(III) with the -COOH of PTA results in an increase in electron cloud density around the PTA's Fe atoms [54]. Compared with the 0.5Fe-P As shown in Figure 4c, compared with that of PTA, the O 1s peak of 2PTA/g-C 3 N 4 moves to a lower binding energy, indicating the increased electron density around the O atoms of PTA and confirming the hydrogen bond interaction between the -NH 2 of g-C 3 N 4 and the -COOH of PTA, combined with the above result of the C-N-H peak shift ( Figure 4b). Interestingly, compared with that of the 2PTA/g-C 3 N 4 , the O 1s peak of 0.5Fe-2PTA/g-C 3 N 4 exhibits a moderate shift to the higher binding energy direction, indicating the decreased electron density around the O atoms of PTA due to the coordination bond interaction between the introduced Fe(III) and the -COOH of PTA. The interaction between the Fe(III) and the -COOH of PTA was also verified by the O 1s peaks of the PTA and 0.5Fe-PTA in Figure S5b, which clearly shows the O 1s peak shift of the PTA from 532.1 eV to 532.4 eV after introducing the Fe(III) [53]. It is worth noting that in Figures 4c and S5b, the O 1s peaks of the 0.5Fe-2PTA/g-C 3 N 4 and 0.5Fe-PTA are higher than those of the 2PTA/g-C 3 N 4 and PTA, respectively, which is due to the -OH of the introduced Fe(III). The Fe 2p spectra of 0.5Fe-2PTA/g-C 3 N 4 (Figure 4d) can be fitted with Fe 2p 3/2 peaks at 710.5 eV and 713.4 eV and the corresponding symmetric peaks of Fe 2p 1/2 at 723.6 eV and 726.4 eV, respectively, while the peaks at 718.7 eV and 731.7 eV are ascribed to the Fe(III) satellite peak [35,48]. These are characteristic of Fe 3+ , further confirming the existence of Fe(III) in 0.5Fe-2PTA/g-C 3 N 4 . To further explore the change of the Fe 2p peaks caused by the interaction between Fe(III) and the -COOH of PTA, the Fe 2p spectra of 0.5Fe-2PTA, 0.5Fe-g-C 3 N 4 and 0.5Fe-2PTA/g-C 3 N 4 were also analyzed. As shown in Figure S5c, compared with those of 0.5Fe-g-C 3 N 4 , the Fe 2p peaks of 0.5Fe-PTA move to a smaller direction, demonstrating that the interaction of Fe(III) with the -COOH of PTA results in an increase in the electron cloud density around the PTA's Fe atoms [54]. Compared with the 0.5Fe-PTA, slightly more obvious shifts to a smaller direction are observed in 0.5Fe-2PTA/g-C 3 N 4 , which may be attributed to the influence of the hydrogen bonds between the -COOH of PTA and the -NH 2 of g-C 3 N 4 .
Based on the above analysis, it can be concluded that the tightly contacted Fe-2PTA/g-C 3 N 4 nanoheterojunctions were successfully developed through the hydrogen bond interaction between the -COOH of PTA and the -NH 2 of g-C 3 N 4 as well as the coordination bond interaction between the Fe(III) and -COOH of PTA. The tight interface contact and the coordinated Fe(III) are expected to facilitate the transfer and separation of photogenerated charges in heterojunctions.

Photophysical and Photochemical Properties
The photogenerated charge separation of the samples were first investigated by photophysical testing with steady-state surface photovoltage spectroscopy (SS-SPS, homebuilt, Harbin, Heilongjiang, China) and photoluminescence (PL, Shimadzu, Kyoto, Japan) measurements. In general, the higher the SS-SPS signal intensity, the more efficient the photogenerated charge separation. As shown in Figure 5a, a relatively low SS-SPS signal is observed on g-C 3 N 4 due to the rapid photogenerated charge recombination, and 2PTA/g-C 3 N 4 shows a more obvious SS-SPS signal than g-C 3 N 4 . In comparison, the SS-SPS signal of 0.5Fe-2PTA/g-C 3 N 4 further increases, indicating its greatly improved photogenerated charge separation [55]. Moreover, as shown in Figure S6, SS-SPS signals first increase and then decrease with an increase in PTA content, and 2PTA/g-C 3 N 4 shows the highest signal, meaning that the shortage of or excessive PTA content is not conducive to photogenerated charge separation. The PL spectra (Figure 5b) also show that 0.5Fe-2PTA/g-C 3 N 4 demonstrates the lowest PL signal compared to 2PTA/g-C 3 N 4 and g-C 3 N 4 , implying that the photogenerated charge recombination is greatly inhibited by constructing nanoheterojunctions with a tight interface contact and coordinating Fe(III) [26]. Furthermore, as shown in Figure S7, xPTA/g-C 3 N 4 exhibits a lower PL signal than g-C 3 N 4 , and when the PTA content reaches 2 wt%, the highest photogenerated charge separation is obtained. Based on the photophysical, photochemical and structural characterization results can be confirmed that constructing g-C3N4 nanoheterojunctions with an appropria amount of PTA and further coordinating Fe(III) facilitate the photogenerated charge tran fer and separation. Additionally, it is anticipated that 0.5Fe-2PTA/g-C3N4 will exhibit To further investigate the photogenerated charge separation in the photocatalytic reaction, the coumarin fluorescent measurement was carried out to detect the amount of formed •OH. A higher fluorescent signal corresponds to a higher amount of •OH, indicating a more efficient charge separation. As shown in Figure 5c, the dramatic fluorescence signal occurs on 2PTA/g-C 3 N 4 and 0.5Fe-2PTA/g-C 3 N 4 compared to g-C 3 N 4 , implying the improved photogenerated charge separation from the construction of nanoheterojunctions. Furthermore, the fluorescence signal of xPTA/g-C 3 N 4 ( Figure S8) is higher than that of g-C 3 N 4 , and when the PTA content reaches 2 wt%, the highest photogenerated charge separation is obtained [56]. The fluorescent intensity result is in agreement with the above SS-SPS and PL results. Figure 5d illustrates the periodic light on/off photocurrent response of the samples. It is obviously observed that the photocurrent intensity of 2PTA/g-C 3 N 4 is higher than that of g-C 3 N 4 , and 0.5Fe-2PTA/g-C 3 N 4 shows the highest photocurrent intensity, confirming the promoted photogenerated charge separation after constructing g-C 3 N 4 nanoheterojunctions with PTA and further coordinating Fe(III). In addition, Nyquist plots of electrochemical impedance spectroscopy (EIS) ( Figure S9) show that whether there is light or not, the arc radius of 0.5Fe-2PTA/g-C 3 N 4 is the smallest compared to the arc radii of 2PTA/g-C 3 N 4 and g-C 3 N 4 , suggesting an increased charge transfer, which helps with the photogenerated charge separation [57,58].
Based on the photophysical, photochemical and structural characterization results, it can be confirmed that constructing g-C 3 N 4 nanoheterojunctions with an appropriate amount of PTA and further coordinating Fe(III) facilitate the photogenerated charge transfer and separation. Additionally, it is anticipated that 0.5Fe-2PTA/g-C 3 N 4 will exhibit an enhanced photocatalytic performance.

Photocatalytic Activities for H 2 Evolution
Photocatalytic activities were examined by detecting the amount of H 2 evolution under illumination with visible light (λ > 420 nm), and the rate constant for H 2 production was analyzed with the kinetic study. Figures 6a and S10 manifest that the visible light photoactivity for H 2 production follows zero-order kinetics. As shown in Figure 6a, g-C 3 N 4 exhibits a meager photocatalytic H 2 evolution activity of 147.1 µmol h −1 g −1 , arising from the poor photogenerated charge separation. In comparison, 2PTA/g-C 3 N 4 exhibits a noticeable increase in the photocatalytic H 2 evolution activity, up to 323.6 µmol h −1 g −1 (~2.2 times that of g-C 3 N 4 ). As expected, the photocatalytic H 2 evolution activity of 0.5Fe-2PTA/g-C 3 N 4 further increases, reaching up to 687.3 µmol h −1 g −1 (~4.6 times of g-C 3 N 4 ) due to the significantly promoted charge separation. We also investigated the effect of PTA and Fe(III) content on the H 2 evolution performance. As shown in Figure S10a,b, with the increase in the content of PTA or Fe(III), the H 2 evolution activity first increases and then decreases. 2PTA/g-C 3 N 4 and 0.5Fe-2PTA/g-C 3 N 4 obtain the highest H 2 evolution activity, respectively, meaning that excessive PTA or Fe(III) will hinder the improvement of the photogenerated charge separation and thus decrease the H 2 evolution activities. The photocatalytic stability of 0.5Fe-2PTA/g-C 3 N 4 was investigated for four cycles of 20 h irradiation. As shown in Figure 6b, the H 2 evolution activity does not exhibit apparent deactivation around any cycle, confirming its considerable stability in the photocatalytic process. Figure 6c shows the wavelength-dependent apparent quantum yield (AQY) values and DRS spectrum of 0.5Fe-2PTA/g-C 3 N 4 . The AQY values of 8.74%, 5.94%, 0.71 % and 0.27% were obtained under the monochromatic irradiation of λ = 365, 405, 470 and 520 nm, respectively, which are similar to the changing tendency of its DRS curves. By comparison, the AQY values present superiority to some other reported g-C 3 N 4 heterojunctions [24,29,[59][60][61]. The photocatalytic H 2 evolution activities at 405 nm and 520 nm excitation wavelengths were further investigated, as shown in Figure 6d. Compared to g-C 3 N 4 , 0.5Fe-2PTA/g-C 3 N 4 shows a prominent improvement for photocatalytic H 2 evolution activities and AQY values at excitation wavelengths of 405 nm and 520 nm. It can also be seen that the photocatalytic H 2 evolution activity of 0.5Fe-2PTA/g-C 3 N 4 at an excitation of 405 nm is better than that at excitation at 520 nm, indicating that the excitation of g-C 3 N 4 is critical for the improved photocatalytic H 2 evolution activities.
the AQY values present superiority to some other reported g-C3N4 heterojunctions [24,29,[59][60][61]. The photocatalytic H2 evolution activities at 405 nm and 520 nm excitation wavelengths were further investigated, as shown in Figure 6d. Compared to g-C3N4, 0.5Fe-2PTA/g-C3N4 shows a prominent improvement for photocatalytic H2 evolution activities and AQY values at excitation wavelengths of 405 nm and 520 nm. It can also be seen that the photocatalytic H2 evolution activity of 0.5Fe-2PTA/g-C3N4 at an excitation of 405 nm is better than that at excitation at 520 nm, indicating that the excitation of g-C3N4 is critical for the improved photocatalytic H2 evolution activities. In order to further confirm the role of -COOH in PTA, H2 evolution measurements of 2PTh/g-C3N4 and 0.5Fe-2PTh/g-C3N4 were carried out for comparison in which there was no -COOH on the main chain of PTh; therefore, there was no hydrogen bond or coordination bond interaction between the PTh and g-C3N4 or Fe(III). It can be seen in Figure S10c that the photocatalytic H2 evolution activity of 2PTh/g-C3N4 only increases by 11% compared to that of g-C3N4, attributed to the poor interface interaction and thus low photogenerated charge transfer and separation between PTh and g-C3N4. Moreover, the photocatalytic H2 evolution activity of 0.5Fe-2PTh/g-C3N4 does not increase further, showing that the introduction of Fe(III) has no positive effect on 2PTh/g-C3N4. By comparison, In order to further confirm the role of -COOH in PTA, H 2 evolution measurements of 2PTh/g-C 3 N 4 and 0.5Fe-2PTh/g-C 3 N 4 were carried out for comparison in which there was no -COOH on the main chain of PTh; therefore, there was no hydrogen bond or coordination bond interaction between the PTh and g-C 3 N 4 or Fe(III). It can be seen in Figure S10c that the photocatalytic H 2 evolution activity of 2PTh/g-C 3 N 4 only increases by 11% compared to that of g-C 3 N 4 , attributed to the poor interface interaction and thus low photogenerated charge transfer and separation between PTh and g-C 3 N 4 . Moreover, the photocatalytic H 2 evolution activity of 0.5Fe-2PTh/g-C 3 N 4 does not increase further, showing that the introduction of Fe(III) has no positive effect on 2PTh/g-C 3 N 4 . By comparison, 2PTA/g-C 3 N 4 and 0.5Fe-2PTA/g-C 3 N 4 exhibit obviously enhanced H 2 evolution activities. This fully demonstrates that the -COOH of PTA interacts with the -NH 2 of g-C 3 N 4 by the hydrogen bond interaction and coordinates with Fe(III), which greatly improves the photocatalytic H 2 evolution activities.

Discussion on Mechanism
To further verify the promoted photogenerated charge transfer and separation of the constructed nanoheterojunctions, the wavelength-dependent fluorescence spectra related to the formed •OH amounts of samples with monochromatic light excitation at 405 and 520 nm were performed, based on the results of Figure 6d. As shown in Figure 7a, under excitation at 405 nm, the strong signal of 0.5Fe-2PTA/g-C 3 N 4 is observed, and its intensity is obviously higher than those of g-C 3 N 4 and 2PTA/g-C 3 N 4 , indicating the significantly improved photogenerated charge separation [8]. And in Figure 7b, under excitation at 520 nm, no response signal is detected in g-C 3 N 4 , and 0.5Fe-2PTA/g-C 3 N 4 presents higher signal intensity than 2PTA/g-C 3 N 4 because in this situation, only PTA is excited and the photogenerated charge transfer occurs from PTA to g-C 3 N 4 . Noticeably, the signal intensities of 0.5Fe-2PTA/g-C 3 N 4 and 2PTA/g-C 3 N 4 under excitation at 520 nm are lower than those under excitation at 405 nm, illustrating that their photocatalytic H 2 evolution activities, and AQY values under excitation at 520 nm are lower than those under excitation at 405 nm in Figure 6d. The single-wavelength photocurrent tests were also employed to investigate the photogenerated charge transfer and separation. Table S1 shows the highest occupied molecular orbital (HOMO) and LUMO position of g-C 3 N 4 and PTA, which were determined from the UV-Vis DRS spectra ( Figure S11) and cyclic voltammetry (CV) curves ( Figure S12). The resulting optical energy gaps (Eg) of g-C 3 N 4 and PTA are 2.70 eV and 1.98 eV, respectively ( Figure S11). A detailed calculation method is shown as Method 1 in the Supporting Information. The optical absorption thresholds of g-C 3 N 4 and PTA are determined to be about 460 and 626 nm, respectively, according to the equation λ = 1240/Eg. As depicted in Figure 7c, with the decrease of the excitation wavelength from 460 to 400 nm, the photocurrent density of g-C 3 N 4 increases gradually. By comparison, 2PTA/g-C 3 N 4 and 0.5Fe-2PTA/g-C 3 N 4 exhibit enhanced photocurrent density from 620 nm to 460 nm, which is due to the visible photosensitization of PTA since in this situation, only PTA is excited [26]. It can be seen that under the irradiation wavelength of 440 nm, the photocurrent density of 2PTA/g-C 3 N 4 increases sharply. According to the HOMO level position of g-C 3 N 4 and the LUMO level position of PTA in Table S1, it can be calculated that 440 nm is the threshold wavelength for the excited high-energy electrons of g-C 3 N 4 to the LUMO of the PTA. Therefore, it can be concluded that the sharp increase in the photocurrent density under 440 nm of irradiation is mainly attributed to the excited, high-level electron transfer from the LUMO of g-C 3 N 4 to the LUMO of PTA, effectively promoting the photogenerated charge separation, and in this process, the slightly higher LUMO of PTA than that of g-C 3 N 4 provides a high-energy electronic platform for the transferred electrons from g-C 3 N 4 . Moreover, an obviously larger photocurrent response is observed in 0.5Fe-2PTA/g-C 3 N 4 than in 2PTA/g-C 3 N 4 , confirming that the coordinated Fe(III) can further facilitate the charge transfer and separation [48,53]. Time-resolved photoluminescence (TR-PL) measurements were also employed to investigate the photogenerated carrier lifetime and dynamic processes. It is accepted that a shortened decay lifetime indicates a more efficient photogenerated electron transfer [62]. As illustrated in Figure 7d, the decay lifetime of g-C 3 N 4 is 2.69 ns, and the decay lifetime is shortened to 2.59 ns in 0.5Fe-2PTA/g-C 3 N 4 , indicating the photogenerated electrons of PTA and g-C 3 N 4 could be extracted in a timely manner by constructing tightly contacted nanoheterojunctions dependent on the hydrogen bond and coordination bond interactions in 0.5Fe-2PTA/g-C 3 N 4 .
In addition, in order to further reveal the role of -COOH in the photogenerated charge transfer and separation of 0.5Fe-2PTA/g-C 3 N 4 , density functional theory (DFT) calculations with a Gaussian B3LYP/6-31G level were conducted to investigate the frontier molecular orbitals of PTA ( Figure S13). It can be observed that the O atoms of -COOH in PTA participate in the HOMO and LUMO of PTA, which is advantageous for the rapid transfer and separation of photogenerated carriers between PTA and g-C 3 N 4 through hydrogen bonds between the -COOH of PTA and the -NH 2 of g-C 3 N 4 [57].
Since Pt, as the cocatalyst, plays a very important role in the photocatalytic H 2 evolution process, we further investigated the dispersion of deposited Pt on 0.5Fe-2PTA/g-C 3 N 4 . As shown in Figure S14a, the TEM image of 0.5Fe-2PTA/g-C 3 N 4 with 1 wt% Pt loaded shows that Pt nanoparticles a few nanometers in size are distributed on the surface. They were also identified in the HRTEM image ( Figure S14b) with a lattice spacing of 0.226 nm, matching the (111) crystal plane of Pt (JCPDS 04-0802). Furthermore, the HAADF-STEM and corresponding EDS mappings ( Figure S14c) clearly show the uniform distribution of Pt on the surface, in addition to C, N, O, S and Fe elements. This is ascribed to the -COOH of PTA and -OH of Fe(III), which are favorable for the evenly dispersed Pt nanoparticles and thus beneficial to generating more active sites during the photocatalysis process and improving the H 2 evolution activities. In addition, in order to further reveal the role of -COOH in the photogenerat charge transfer and separation of 0.5Fe-2PTA/g-C3N4, density functional theory (DFT) c culations with a Gaussian B3LYP/6-31G level were conducted to investigate the front molecular orbitals of PTA ( Figure S13). It can be observed that the O atoms of -COOH PTA participate in the HOMO and LUMO of PTA, which is advantageous for the rap transfer and separation of photogenerated carriers between PTA and g-C3N4 through h drogen bonds between the -COOH of PTA and the -NH2 of g-C3N4 [57].
Since Pt, as the cocatalyst, plays a very important role in the photocatalytic H2 evo tion process, we further investigated the dispersion of deposited Pt on 0.5Fe-2PTA/g-C3N As shown in Figure S14a, the TEM image of 0.5Fe-2PTA/g-C3N4 with 1 wt% Pt load shows that Pt nanoparticles a few nanometers in size are distributed on the surface. Th were also identified in the HRTEM image ( Figure S14b) with a lattice spacing of 0.226 n matching the (111) crystal plane of Pt (JCPDS 04-0802). Furthermore, the HAADF-STE and corresponding EDS mappings ( Figure S14c) clearly show the uniform distribution Pt on the surface, in addition to C, N, O, S and Fe elements. This is ascribed to the -COO of PTA and -OH of Fe(III), which are favorable for the evenly dispersed Pt nanopartic Based on the above analysis, the possible mechanism of photogenerated charge transfer and separation was proposed in Figures 8 and S14. It has been confirmed that 440 nm is the threshold wavelength for the high-level electron transfer ( Figure 7c). As shown in Figure 8, under irradiation with visible light below 440 nm, although both the g-C 3 N 4 and PTA in the nanoheterojunctions are excited simultaneously, in the nanoheterojunctions with g-C 3 N 4 as the main part, the photogenerated charge separation enhancement mainly results from the high-energy electron transfer from the LUMO of g-C 3 N 4 to the LUMO of PTA and the continuous transfer to the coordinated Fe(III) with -OH and a Pt cocatalyst. As shown in Figure S15, when the nanoheterojunctions are irradiated by visible light in a wavelength range from 460 nm to 620 nm, only PTA is excited: therefore, the photogenerated electrons of PTA transfer from the LUMO of PTA to the LUMO of g-C 3 N 4. In this situation, PTA, as a photosensitizer, improves the utilization range of visible light. wavelength range from 460 nm to 620 nm, only PTA is excited: therefore, the photogen ated electrons of PTA transfer from the LUMO of PTA to the LUMO of g-C3N4. In t situation, PTA, as a photosensitizer, improves the utilization range of visible light.

Conclusions
In summary, using the in situ photopolymerization and impregnation methods, have successfully constructed Fe(III)-coordinated poly(3-thiophenecarboxylic acid) C3N4 nanoheterojunctions with tight interface contact for visible-light-driven H2 evo tion. Compared to g-C3N4, the ratio-optimized 0.5Fe-2PTA/g-C3N4 exhibits a ~4.6-fold hancement of visible-light photocatalytic H2 evolution activity, attributed to the sign cantly promoted photogenerated charge transfer and separation and the expanded v ble-light response. It is proved that the tight interface contact, which is dependent on hydrogen bond interaction between the -COOH of PTA and the -NH2 of g-C3N4 and coordination bond interaction between Fe(III) and the -COOH of PTA, plays a key role improving the photogenerated charge transfer and separation. It is also confirmed t the enhanced photoactivity is mainly ascribed to the high-energy electron transfer fr the LUMO of g-C3N4 to the LUMO of PTA and the continuous transfer to the coordina Fe(III) with -OH and a Pt cocatalyst in the nanoheterojunctions with g-C3N4 as the m part. This work not only provides a very feasible reference to design and synthesize e cient g-C3N4 heterojunction photocatalysts for visible-light-driven energy evolution b also helps us to understand the charge transfer and separation mechanisms in detail.

Conclusions
In summary, using the in situ photopolymerization and impregnation methods, we have successfully constructed Fe(III)-coordinated poly(3-thiophenecarboxylic acid)/g-C 3 N 4 nanoheterojunctions with tight interface contact for visible-light-driven H 2 evolution. Compared to g-C 3 N 4 , the ratio-optimized 0.5Fe-2PTA/g-C 3 N 4 exhibits a~4.6-fold enhancement of visible-light photocatalytic H 2 evolution activity, attributed to the significantly promoted photogenerated charge transfer and separation and the expanded visible-light response. It is proved that the tight interface contact, which is dependent on the hydrogen bond interaction between the -COOH of PTA and the -NH 2 of g-C 3 N 4 and the coordination bond interaction between Fe(III) and the -COOH of PTA, plays a key role in improving the photogenerated charge transfer and separation. It is also confirmed that the enhanced photoactivity is mainly ascribed to the high-energy electron transfer from the LUMO of g-C 3 N 4 to the LUMO of PTA and the continuous transfer to the coordinated Fe(III) with -OH and a Pt cocatalyst in the nanoheterojunctions with g-C 3 N 4 as the main part. This work not only provides a very feasible reference to design and synthesize efficient g-C 3 N 4 heterojunction photocatalysts for visible-light-driven energy evolution but also helps us to understand the charge transfer and separation mechanisms in detail.