MoxC Heterostructures as Efficient Cocatalysts in Robust MoxC/g-C3N4 Nanocomposites for Photocatalytic H2 Production from Ethanol

In this work, we studied new materials free of noble metals that are active in photocatalytic H2 generation from ethanol aqueous solutions (EtOHaq), which can be obtained from biomass. MoxC/g-C3N4 photocatalysts containing hexagonal (hcp) Mo2C and/or cubic (fcc) MoC nanoparticles on g-C3N4 nanosheets were prepared, characterized, and evaluated for photocatalytic hydrogen production from EtOHaq (25% v/v). Tailored MoxC/g-C3N4 nanocomposites with MoxC crystallite sizes in the 4–37 nm range were prepared by treatment with ultrasound of dispersions containing MoxC and g-C3N4 nanosheets, formerly synthesized. The characterization of the resulting nanocomposites, MoxC/g-C3N4, by different techniques, including photoelectrochemical measurements, allowed us to relate the photocatalytic performance of materials with the characteristics of the MoxC phase integrated onto g-C3N4. The samples containing smaller hcp Mo2C crystallites showed better photocatalytic performance. The most performant nanocomposite contained nanoparticles of both hcp Mo2C and fcc MoC and produced 27.9 mmol H2 g–1 Mo; this sample showed the lowest recombination of photogenerated charges, the highest photocurrent response, and the lowest electron transfer resistance, which can be related to the presence of MoC-Mo2C heterojunctions. Moreover, this material allows for easy reusability. This work provides new insights for future research on noble-metal-free g-C3N4-based photocatalysts.


INTRODUCTION
Nowadays, the large-scale industrial production of H 2 is captive to fossil fuels.Thus, new ways to produce H 2 are of current interest in both the energy and environmental contexts.−4 Photocatalysts are usually based on inorganic semiconductor materials, and the direct use of solar radiation is highly attractive when the photocatalyst is able to absorb visible light.−11 On the other hand, the main concern related to the efficiency of photocatalysts is the usually fast rate of recombination of the photogenerated charges, electrons (e − ), and holes (h + ). 12 Although the most common strategy that is used for improving h + /e − charge separation is the addition of noble metals to the surface of the semiconductor, other compounds such as transition metal oxides, sulfides, or carbides, have been also proposed as efficient cocatalysts. 13,14−26 The use of molybdenum carbides in alternative hydrogen production, has been mainly focused on their use as electrocatalysts. 14However, an increasing interest exists in the use of molybdenum carbide-based systems as cocatalysts in the photoproduction of H 2 . 14−35 Mo−Mo 2 C/g-C 3 N 4 and Co-doped Mo− Mo 2 C/g-C 3 N 4 , 31,35 Mo 2 C@C/g-C 3 N 4 heterostructures, 32,33 and rod-like g-C 3 N 4 decorated with Mo 2 C, 34 have been studied as photocatalysts in H 2 generation; in every case, triethanolamine (TEOA) solution was employed as sacrificial electron donor in the photocatalytic test.−35 On the other hand, it has been reported that the presence of MoC/Mo 2 C heterostructures has a positive effect in the electrocatalytic behavior of molybdenum carbide-based catalysts for the H 2 evolution reaction. 36Moreover, it has been proposed that, for a given material, there is a relationship between its performance in the electrocatalytic H 2 evolution and its efficiency as a cocatalyst in the photocatalytic H 2 production. 35his background and the advantage of using ethanol, which is currently produced and used as a biofuel around the world, led us to an in-depth study of photocatalysts based on Mo x C/ g-C 3 N 4 (Mo x C = Mo 2 C and/or MoC) for H 2 production using EtOH aq and visible light.−39 In this work, we report the preparation and characterization of new tailored Mo x C/g-C 3 N 4 photocatalysts, containing hcp Mo 2 C and/or fcc MoC as cocatalysts with different crystallite sizes.Hydrogen production is related to the photoelectrochemical properties of the nanocomposites, which in turn depend on the crystalline phases of molybdenum carbide, including charge recombination, electron transfer resistance, and the photocurrent response of different photocatalysts.

Synthesis of Mo
x C. Different Mo x C phases were synthesized based on a recently described sol−gel method using MoCl 5 (reagent grade, 95%) and 4,5-dicyanoimidazole (DI) (reagent grade, 99%) from Sigma-Aldrich, as Mo and C sources, respectively. 37,38pecifically, the gel formed by the addition of MoCl 5 (5.6 mmol) and DI (2.8 mmol) to 15 mL of ethanol (>99.999%HPLC, Sigma-Aldrich), was treated at different temperatures under an Ar flow (99.999%,Linde) (T = 700, 800, or 900 °C) in a tubular furnace.The materials were labeled Mo x CT, where T is the temperature used in the treatment.The preparation of Mo x C1100 was accomplished using a similar method, with MoCl 5 (5.6 mmol), DI (5.6 mmol), and a thermal treatment of 1100 °C.

Preparation of g-C 3 N 4 Nanosheets.
For the preparation of g-C 3 N 4 nanosheets, the thermal polymerization of melamine was accomplished as follows: 40 melamine (>99%, Alfa Aesar) was calcined at 5 °C/min up to 520 °C (4 h); the yellowish material was ground to powder, and the calcination process was repeated.

Integration of Mo
x CT onto g-C 3 N 4 .For the preparation of Mo x CT/g-C 3 N 4 , dispersions in ethanol with the appropriate amounts of Mo x CT and g-C 3 N 4 , to obtain about 3 wt% Mo, were treated under ultrasound (SONICS VCX 500) at 20 °C and 250 W for 1 h; then, ethanol was eliminated by careful evaporation with stirring at 50 °C.Moreover, using a similar procedure, commercial Alfa-Aesar Mo 2 C (99.5% metal basis, hcp Mo 2 C, average crystallite size d̅ = 37 nm), was used for the preparation of the Mo 2 C-comm/g-C 3 N 4 photocatalyst.Table 1 lists all of the photocatalysts prepared and studied in this work.
2.4.Characterization of Photocatalysts.Inductively coupled plasma atomic emission spectrometry (ICP-AES) was used to determine the Mo content using a PerkinElmer Optima 3200RL instrument.
A Micromeritics Tristar II 3020 instrument was used to record the N 2 adsorption−desorption isotherms, which were determined at −196 °C after degasification at 250 °C under an Ar flow.Multipoint Brunauer−Emmett−Teller (BET) analysis of the isotherms was used to calculate the specific surface area (S BET ).The desorption isotherms were used for the determination of the pore size distribution by the method of Barrett−Joyner−Halenda (BJH).
Powder X-ray diffraction (XRD) patterns were recorded using a Bragg−Brentano powder diffractometer (PANalytical X'Pert PRO MPD), with Cu Kα radiation (λ = 1.5418Å) from 2θ = 4 to 100°.Peak indexation and phase identification were performed with the aid of the ICDD Powder Diffraction File (PDF). 41Accurate peak positions, area intensities, and full width at half-maximum (fwhm) were obtained after full profile analysis.The crystallite sizes were calculated from the fwhm using the Scherrer equation. 42Semiquantitative phase analysis, in some binary nanocrystalline samples, was performed from the accurate area intensities and using the reference intensity ratios in the corresponding PDF file. 43ourier-transform infrared spectroscopy (FTIR) analysis was conducted by using pellets of the KBr-diluted samples on a Thermo Nicolet 5700 FTIR apparatus.
Transmission and high-resolution electron microscopy (TEM-HRTEM) and energy-dispersive X-ray (EDX) spectroscopy with elemental mapping analysis were performed using a JEOL JEM-2100 instrument operating at 200 kV.
X-ray photoelectron spectroscopy (XPS) was conducted using a PerkinElmer PHI-5500 Multitechnique System with Al Kα radiation (1486.6 eV).The binding energy (BE) values were determined using the C 1s peak at 284.8 eV, which was previously ascertained using Au as a reference.MultiPak XPS software was used to deconvolute the XPS signals.
UV−vis diffuse reflectance spectra (UV−vis DRS) were acquired using a PerkinElmer Lambda 950 UV/vis Spectrometer; BaSO 4 was used as a reference.The Kubelka−Munk model was used to determine the band gap values.
Photoluminescence (PL) spectroscopy measurements were performed at room temperature by using a Kimmon IK Series He−Cd CW laser (325 nm, 40 mW).Fluorescence was dispersed through a SpectraPro 2750 (focal length of 750 mm) f/9.8 monochromator, detected with a Hamamatsu H8259-02 photomultiplier, and amplified Mo content from chemical analysis (ICP-AES), BET surface area, position of the 002 XRD peak of g-C 3 N 4 , Mo x C phase, crystallite size, and band gap values.b Calculated from the full profile analysis of the XRD pattern (Figure S6 43 ).using a Stanford Research System SR830 DSP Lock-in amplifier.A 360 nm filter was used for the stray light, and the emission spectra were corrected using the optical transfer function of the PL setup.
For photoelectrochemical characterization of samples, electrochemical impedance spectroscopy (EIS) and transient photocurrent determination were performed.The system consisted of a computercontrolled potentiostat (VMP3, BioLogic Science Instruments) with an undivided three-electrode cell; a Pt wire was used as the counter electrode, Ag/AgCl (3 M KCl) as the reference, the photocatalyst (1 cm 2 geometric area) as the working electrode; and an aqueous solution of Na 2 SO 4 (0.5 M) as the electrolyte.A150 W AM 1.5G solar simulator (Solar Light Co., 16S-300-002 v 4.0) with an incident light intensity of 1 sun (100 mW cm −2 ) was used to perform the measurements under illumination.
2.5.Photocatalytic Experiments.Figure S1 shows a schematic diagram of the system used for the photocatalytic tests.It contained a jacketed reactor of 300 mL, operated under continuous gas flow, and was equipped with a condenser, kept at −15 °C at the outlet. 44A broad-spectrum commercial (ACE-Hanovia) Hg lamp was used (Figure S1), which was immersed in the solution inside a watercooled jacket that served as a UV cut-off filter (λ > 385 nm).The experiments were performed at atmospheric pressure and 20 °C.Before the photocatalytic test, the photocatalyst (300 mg) was dried at 100 °C, and EtOH aq (250 mL, 25% v/v) was purged with Ar; and the corresponding suspension was stirred in the reactor for 30 min under dark and N 2 /Ar flow (99.999%N 2 was used as the internal standard) and then irradiated.No products were detected in the dark.After 10 min of light on, the evolved gaseous products were periodically analyzed online using a gas microchromatograph Varian CP-4900, equipped with micro-TCD detectors (detection limit for H 2 , 50 ppm) and two columns (10 m PPQ, He carrier; 10 m molecular sieve (5 Å), Ar carrier).
After the test (4 h), the mixture was filtered, and the solution was analyzed by gas chromatography using Bruker 450 GC equipment with an FID detector and a CP-Sil 8 CB capillary column (30 m × 0.25 mm).
A reusability test was carried out with the most active photocatalyst; the used sample was removed from the liquid reaction mixture by filtering, washed with ethanol, and tested again.Moreover, in a separate cyclic experiment, the light was switched off after 1 h, and then after 0.5 h in dark conditions, the light was switched on for 1 h.

Structural and Chemical Characterization.
As stated in the Experimental Section, the Mo x CT/g-C 3 N 4 catalysts studied in this work were prepared by ultrasonic treatment of a suspension containing previously synthesized molybdenum carbide nanoparticles and g-C 3 N 4 nanosheets.XRD patterns of the initial Mo x C800 and Mo x C900 (Figure S2) show the main presence of relatively narrow peaks that could be perfectly indexed with the hexagonal close-packed, hcp, P6 3 /mmc, structure of Mo 2 C (PDF 04-014-1517).The average crystallite size of Mo 2 C was d̅ = 22 and 30 nm in Mo x C800 (from now Mo 2 C800) and Mo x C900 (from now Mo 2 C900), respectively.On the other hand, in the XRD pattern of Mo x C1100 (Figure S2), mainly 6 very wide peaks are observed that could be in principle be well indexed as the 111, 200, 220, 311, 222, and 400 reflections of the cubic closepacked, fcc, Fm3̅ m, the structure of MoC (PDF 04-003-1480) with d̅ = 4 nm; for easy identification of this sample, from this point on it will be labeled MoC1100.Meanwhile, the XRD analysis of Mo x C700 (Figure S2) shows both the presence of very wide peaks of the fcc structure of MoC and less wide peaks of the hcp structure of Mo 2 C.
The XRD pattern of g-C 3 N 4 prepared in this work shows a main peak at 2θ = 27.7°and a peak at 2θ = 13.0°with a much lower intensity (Figure 1).The peak at 2θ = 27.7°isdue to 002 reflection of the (001) interlayer stacking; the position of this peak indicates a slightly smaller interlayer distance with respect to bulk g-C 3 N 4 , which shows the 002 XRD peak at 27.34°. 40,45The low intensity peak at 2θ = 13.0°isattributed to the in-plane structural packing motifs. 46he FTIR spectrum of g-C 3 N 4 (Figure S3) agrees with that expected for this material. 40,45The characteristic band of the breathing mode of s-triazine rings can be seen at 807 cm −1 ; bands related to C�N and C−N stretchings in the heterocycles are visible in the 1800−900 cm −1 zone. 40,45All of these FTIR features can also be seen in the spectra of all nanocomposites prepared (Figure S3).
All photocatalysts showed type IV N 2 adsorption− desorption isotherms with H3 hysteresis loops (Figure S4). 47he S BET values of 29−39 m 2 g −1 (Table 1), and wide pore size distributions in the range of meso-and macropores were found (Figure S5).
Figure 1 shows the XRD patterns of Mo x CT/g-C 3 N 4 nanocomposites.After the integration of Mo x CT onto g-C 3 N 4 , the photocatalysts always retained the initial Mo x C crystalline phases and g-C 3 N 4 nanosheets used in their preparation (Figure 1).
Table 1 shows the obtained crystallite sizes of the Mo x C phases in the photocatalysts.The semiquantitative Mo x C phase analysis of the Mo x C700/g-C 3 N 4 sample results in hcp Mo 2 C (14%) and fcc MoC (86%). 43exagonal Mo 2 C (11−31 nm crystallite size) was identified in the Mo x CT/g-C 3 N 4 (T = 700−900 °C) nanocomposites; the higher the temperature used in the preparation of Mo x CT, the larger the Mo 2 C crystallite size.Mo 2 C-comm/g-C 3 N 4 showed the largest hcp Mo 2 C crystallite size (37 nm).On the other hand, the fcc MoC phase, which was determined in Mo x C700/g-C 3 N 4 and MoC1100/g-C 3 N 4 , showed a similar crystallite size (4 nm) in both nanocomposites.
Figure 3 shows TEM and HRTEM images of Mo x C700/g-C 3 N 4 .In this case, the nanoparticles of Mo x C with two domains of particle size, with mean sizes of 4.3 and 13.0 nm, can be seen (Figure 3B).Moreover, the presence of hcp Mo 2 C and fcc MoC in close proximity was determined by HRTEM (Figure 3C).These results are in good agreement with those of the XRD analysis.Moreover, EDX analysis showed that Mo was homogeneously distributed along the photocatalyst (Figure 3D).
The surface characteristics of the photocatalysts were analyzed by XPS.The XP C 1s spectrum corresponding to g-C 3 N 4 (Figure S7) shows, besides the adventitious band due to C−C at 284.8 eV, a main band at about 288.1 eV, characteristic of bound C in the heterocycles (C−N�C). 40,48n addition, a small C 1s component at a lower BE (283.3−283.8 eV), characteristic of carbides can be identified in the XP spectra of Mo x C-containing photocatalysts (Figure S7). 24,37,38n all cases, the broadness and asymmetry of the band located at the highest BE are indicative of the presence of O-bound (C−O, C�O) surface species. 24,40,49The N 1s spectrum of g-C 3 N 4 (Figure S7) shows a main component assigned to sp 2bonded N in the heterocycles at about 398.6 eV, and components at higher BEs of 399.4,401.0, and 404.7 eV.The signal at 399.4 eV can be attributed to the presence of tertiary nitrogen (N-(C) 3 ).On the other hand, amino functional groups (C−N−H), arising from a defective condensation of heptazine substructures, could be responsible for the component at 401.0 eV; the very low intense peak at 404.7 eV is attributed to the charging effects or positive charge localization in the heterocycles. 48,50,51The N 1s spectra of the  nanocomposites were similar to those of g-C 3 N 4 (Figure S7).On the other hand, the presence of a small amount of surface oxygen-containing species could be evidenced in all cases (Figure S8); the asymmetric broad O 1s band with a maximum at 532.6 eV in the XP spectrum of g-C 3 N 4 could be assigned to adsorbed H 2 O and surface species with C�O bonds. 40olybdenum oxide and/or oxycarbide species could contribute to the O 1s band in the XP spectra of Mo x C-containing photocatalysts, whose maxima are located at a lower BE than that of g-C 3 N 4 . 24,49Finally, Figure 4 shows the Mo 3d core level spectra of nanocomposites, which were deconvoluted fixing the Mo 3d 5/2 /Mo 3d 3/2 intensity ratios of 1.5 and 3.1 eV as the value of orbital splitting. 52The Mo 3d 5/2 component at 228.4−228.7 eV indicates the presence of surface carbides and is attributed to Mo 2 C and/or derived oxycarbide species. 24,37,49,53The other Mo 3d 5/2 components at higher BEs can be assigned to the Mo 4+ , Mo 5+ , and Mo 6+ surface species, which could be related to MoC and/or different molybdenum oxycarbide and oxide species. 37,39,53The samples Mo x C700/g-C 3 N 4 , Mo x C800/g-C 3 N 4 , and Mo x C900/g-C 3 N 4 showed clear Mo 3d 5/2 components at a BE lower than 231 eV, and the corresponding molybdenum surface species were about 23% of the total Mo n+ surface species.Although the Mo 3d spectrum of MoC1100/g-C 3 N 4 could not be properly deconvoluted, the existence of surface molybdenum carbide and oxide species could also be inferred in this case (Figure 4).In all cases, the contact of samples with ambient air could produce oxycarbide and oxide species. 24.2.Photoelectrochemical Features and Photocatalytic Behavior.The optical properties of the photocatalysts were analyzed by UV−vis eliminate DRS and PL spectroscopy.Figure S9 shows the UV−vis DRS spectra and the corresponding Tauc plots used in the determination of the band gap values, following the approach reported in ref 54.In all cases, light absorption edges occurred in the visible region (Table 1).The band gap determined for g-C 3 N 4 is 2.75 eV, and all of the Mo x CT/g-C 3 N 4 nanocomposites showed only slightly higher band gap values (2.77−2.79eV).
The recombination process of photogenerated (e − /h + ) pairs in the photocatalysts was studied by PL analysis.Figure 5 shows the PL spectra of the nanocomposites compared to that of g-C 3 N 4 .We observed an emission peak in the visible region for all samples, a maximum at 465 nm for g-C 3 N 4 , and a slight shift of the emission maxima toward higher energies for the Mo x C-containing photocatalysts.These findings are consistent with the UV−vis DRS results, which indicate a slight increase in the band gap energy for Mo x CT/g-C 3 N 4 than pristine g-C 3 N 4 .Interestingly, the intensity of the PL spectrum of g-C 3 N 4 is always higher than that of nanocomposites (Figure 5), indicating that the presence of Mo x C decreases the e − /h + pair recombination rate, thus leading to a decrease in the emission intensity.The trend in the intensity of the PL emission band (Figure 5) suggests that hcp Mo 2 C is more effective in reducing charge recombination in g-C 3 N 4 than fcc MoC.Moreover, it seems that recombination is less favored when the Mo 2 C crystallite size decreases and when hcp Mo 2 C and fcc MoC are in close proximity over g-C 3 N 4 .A low recombination rate of the photogenerated (e − /h + ) pairs and a faster electron transfer rate can improve the photocatalytic properties.
To further evaluate the photocharge separation and electron transfer properties of the photocatalysts, electrochemical impedance spectra, and transient photocurrent response were measured, as described in the Experimental Section.
Figure 6 shows the impedance Nyquist plots for the g-C 3 N 4 and Mo x CT/g-C 3 N 4 photocatalysts under dark (A) and illumination (B) conditions.As can be seen, in both cases, g-C 3 N 4 shows a larger arc radius than the Mo x C-containing nanocomposites, indicating that the presence of Mo x C on the g-C 3 N 4 nanosheets reduces the electron transfer resistance in the dark and under irradiation.Moreover, in all cases, the arc radius in the dark is larger than that under irradiation, as illustrated in Figure S10 for the samples with the largest and the smallest arc radii, g-C 3 N 4 and Mo x C700/g-C 3 N 4 , respectively.This indicates the effectiveness of irradiation in decreasing the barrier of electron transfer in these materials.The Nyquist arc radius follows a trend similar to that of the intensity of the PL spectra shown above; the lower the electron transfer barrier, the higher the efficiency of charge separation.
The results of the transient photocurrent responses are shown in Figure 7.The Mo x C-containing nanocomposites   showed a larger photocurrent density than that of g-C 3 N 4 .Mo x C700/g-C 3 N 4 showed the highest photocurrent density.The presence of Mo x C as a cocatalyst slows down the recombination rate of photogenerated e − /h + , decreases the barrier for electron transport, and accordlingly can favor the further transfer of electrons for proton reduction.
As stated above, the photocatalytic behavior in H 2 generation from EtOH aq (25% v/v) was analyzed under visible light (λ > 385 nm) for all of the materials prepared, including pristine g-C 3 N 4 , and a blank test without the photocatalyst was also carried out.Under the experimental conditions used in this work, g-C 3 N 4 and the blank test produced a negligible amount of H 2 .On the other hand, all nanocomposites containing molybdenum carbide species were active in photocatalytic H 2 generation; the H 2 production per gram of catalyst over time is depicted in Figure 8.Besides H 2 , minor amounts of CO 2 were produced; moreover, mainly 2,3butanediol was found in the liquid phase, and no O 2 was detected.These results point out that photoreforming of ethanol was not accomplished.Moreover, the absence of CH 4 and CO as products in the gas phase suggests that under our experimental conditions and withMo x C/g-C 3 N 4 photocatalysts, the cleavage of the C−C bond is not favored in the ethanol phototransformation. 1,2,4 The oxidation of ethanol could proceed with the generated hole (h + ), and the αhydroxyethyl radicals ( • CH(OH)CH 3 ) be formed: 38,55,56 Then, the coupling of two initially formed α-hydroxyethyl radicals ( • CH(OH)CH 3 ) could be proposed for the formation of 2,3-butanediol: (2) According to the characterization results given above, the molybdenum carbide species on g-C 3 N 4 mainly facilitate electronic transfer, producing H 2 and decreasing e − /h + recombination in g-C 3 N 4 .This is illustrated in Figure S11 for Mo x C700/g-C 3 N 4 .
Figure 8 shows the hydrogen production per gram of the Mo x CT/g-C 3 N 4 photocatalyst; a similar trend is found when H 2 production is referred to as a gram of Mo (Figure S12).The photocatalytic performance can be straightforwardly related to the photoelectrochemical characteristics of the samples: the intensity of the PL emission peak, EIS values, and transient photocurrent measurements.The smaller the size of the hcp Mo 2 C particles in the nanocomposite, the better the photocatalytic performance.Moreover, a positive effect of the close proximity of the hcp Mo 2 C and fcc MoC phases is found.Mo x C700/g-C 3 N 4 with both hcp Mo 2 C (11 nm) and fcc MoC (4 nm) showed the best photocatalytic behavior.The presence of heterojunctions MoC-Mo 2 C could improve the photocatalytic behavior, as has been demonstrated for WC-Mo 2 C/ TiO 2 photocatalysts with WC-Mo 2 C heterojunctions. 58In fact, as stated in the Introduction section, for Mo x C-based electrocatalysts, the presence of MoC/Mo 2 C heterostructures was demonstrated to improve the H 2 evolution reaction, 36 and an improved photocatalytic behavior for H 2 production could be expected, 35 which is in line with the results presented here.
In order to know the reusability of the most performant photocatalyst, a new photocatalytic test was carried out with the used Mo x C700/g-C 3 N 4 (Figure 9), and a similar H 2 production for both fresh and used catalysts was observed.Figure 9 also shows the reproducibility of the photocatalytic behavior of fresh Mo x C700/g-C 3 N 4 when the light switch off/ switch on procedure was employed.In all cases, the standard deviation was only up to 2% of the values originally obtained.The reused Mo x C700/g-C 3 N 4 was characterized by FTIR and XRD, and similar results were obtained for the fresh photocatalysts (Figures S13 and S14).These results show the stability of Mo x C700/g-C 3 N 4 , which can be reused without loss of its properties.

CONCLUSIONS
The preparation method led to tailored Mo x CT/g-C 3 N 4 photocatalysts with hcp Mo 2 C and/or fcc MoC nanoparticles of different sizes supported on g-C 3 N 4 nanosheets.The separately prepared Mo x C cocatalysts were incorporated into g-C 3 N 4 using ultrasound, which maintained the characteristics  of both Mo x C nanoparticles and g-C 3 N 4 nanosheets formerly synthesized.
The photocatalytic behavior of Mo x CT/g-C 3 N 4 in hydrogen generation from EtOH aq under visible light is related to the rate of recombination of photogenerated charges, electron transfer resistance, and photocurrent response of the prepared nanocomposites.These properties depend on the characteristics of the Mo x C cocatalyst.Photocatalysts containing hcp Mo 2 C were more effective than that containing only fcc MoC; the smaller the crystallite size of hcp Mo 2 C, the better was the photocatalytic performance of Mo x CT/g-C 3 N 4 .The best photocatalytic results were obtained for Mo x C700/g-C 3 N 4 , which presented hexagonal Mo 2 C and cubic MoC, the lowest rate of charge recombination and electron transfer resistance, and the highest photocurrent response; about 7 mmol of H 2 g Mo −1 h −1 were produced with this photocatalyst under the conditions used.The improved photocatalytic behavior of Mo x C700/g-C 3 N 4 could be related to the presence of MoC-Mo 2 C heterojunctions, which could enhance the photoelectrochemical properties of the nanocomposites and therefore photocatalytic H 2 generation.Moreover, the reusability of Mo x C700/g-C 3 N 4 is demonstrated.
Additional details related to characterization methods and characterization and photocatalytic results are included (PDF)

Figure 4 .
Figure 4. Mo 3d core level spectra of the nanocomposites.

Figure 6 .
Figure 6.EIS Nyquist plots of nanocomposites and g-C 3 N 4 : (A) under dark conditions and (B) with irradiation.

Figure 8 .
Figure 8. H 2 generated per gram of photocatalyst as a function of irradiation time; results from g-C 3 N 4 and the blank test are also included.Reaction conditions: 25% v/v ethanol aqueous solution and T = 20 °C.

Figure 9 .
Figure 9. Different photocatalytic tests for H 2 production over fresh and used Mo x C700/g-C 3 N 4 .In test 2, the light was switched off after 1 h; then, after 0.5 h in dark conditions, the light was switched on.Reaction conditions: 25% v/v ethanol aqueous solution and T = 20 °C.

Table 1 .
Several Characteristics of Photocatalysts a