Effect of Ruthenium Modiﬁcation of g-C 3 N 4 in the Visible-Light-Driven Photocatalytic Reduction of Cr(VI)

: Graphitic carbon nitride (g-C 3 N 4 ) is a promising heterogeneous photocatalyst in the visible range. It can be used, among others, for reductive conversion of the toxic hexavalent chromium occurring in various wastewaters. Its photocatalytic efﬁciency, however, has to be improved, which can be realized by modiﬁcation with different dopants or co-catalysts forming heterojunctions. In our work, ruthenium-modiﬁed g-C 3 N 4 has been prepared by ultrasonic impregnation of the pristine g-C 3 N 4 , which was synthesized from thiourea. The morphology, microstructure, and optical properties of the photocatalysts were characterized by XRD, SEM, FT-IR, TEM, XPS, and DRS. Their compositions were analyzed by EDS and XPS measurements, indicating 0.5% and 1.4% Ru, due to the different penetrating depths. XPS study showed mainly +2 for the oxidation state of Ru. DRS analysis indicated a slight change in both the CB (from − 1.14 to − 1.22 eV) and the VB (from 1.49 to 1.56 eV) energies of Ru/g-C 3 N 4 , compared to those of g-C 3 N 4 . The photocatalytic Cr(VI) reduction efﬁcacy increased from 50.1 to 96.8%. Low pH (=2) was preferred for the photocatalytic Cr(VI) reduction due to the favorable surface charge and E(Cr(VI)/Cr(III)) redox potential. Ru modiﬁcation proved to be promising for improving the photocatalytic performance of g-C 3 N 4 .


Introduction
Heterogeneous photocatalysis became an intensively developing field of science in past decades because it offers good possibilities for the removal of various types of pollutants [1]. In several cases, for this purpose, utilization of solar radiation can also be realized by photoactive semiconductors. Unfortunately, a significant part of the stable and efficient heterogeneous photocatalysts are white (such as TiO 2 , SnO 2 , ZnO), and thus they can only be excited in the UV range [2], which represents only a low fraction of the solar light. Hence, more and more interest has been attracted by colored semiconductors, a considerable part of which is organic. The advantage of organic semiconductors (OSCs) is that their basic structure is metal-free, which is important from the viewpoint of environmental protection, and they are relatively easy to prepare in rather mild experimental conditions. One of the most promising OSCs is graphitic carbon nitride (g-C 3 N 4 ), which is a conjugated polymer [3]. It is suitable for efficient utilization of the longer-wavelength photons due to Ru/C indicates that it has a quasi-graphitic structure. The reflection at 13.08° is indexed as (100) peak that arises from the in-plane ordering of tri-s-triazine attributed to units of g-C3N4. This means that the loading of Ru did not change the basic structure of g-C3N4. The absence of an apparent characteristic peak of Ru on Ru/g-C3N4 indicates that Ru has a small particle size, low loading, and good dispersion on the g-C3N4 surface. The decreased intensity of the characteristic peaks in the XRD patterns of Ru/g-C3N4 may be attributed to a reduced layer thickness caused by Ru doping [27].

UV-Vis Spectroscopic Study
Absorption spectroscopy is used to determine the optical properties of the synthesized samples, which is an important factor for photocatalysts. Figure 2 shows the DR/UV-Vis diffuse reflectance spectra and the Kubelka-Munk plot [28] of the sample's Ru/g-C3N4 and g-C3N4. The significant decrease in the reflection upon Ru doping ( Figure  2a) can be attributed to the enhanced light absorption, which is in accordance with the much darker color. Absorption in the visible light range (from 400 to 750 nm) is an important condition for the photocatalytic activity of g-C3N4 under visible light. When doping Ru NPs onto the

UV-Vis Spectroscopic Study
Absorption spectroscopy is used to determine the optical properties of the synthesized samples, which is an important factor for photocatalysts. Figure 2 shows the DR/UV-Vis diffuse reflectance spectra and the Kubelka-Munk plot [28] of the sample's Ru/g-C 3 N 4 and g-C 3 N 4 . The significant decrease in the reflection upon Ru doping (Figure 2a) can be attributed to the enhanced light absorption, which is in accordance with the much darker color.
Ru/C indicates that it has a quasi-graphitic structure. The reflection at 13.08° is inde (100) peak that arises from the in-plane ordering of tri-s-triazine attributed to unit C3N4. This means that the loading of Ru did not change the basic structure of g-C3N absence of an apparent characteristic peak of Ru on Ru/g-C3N4 indicates that Ru small particle size, low loading, and good dispersion on the g-C3N4 surface. The dec intensity of the characteristic peaks in the XRD patterns of Ru/g-C3N4 may be attribu a reduced layer thickness caused by Ru doping [27].

UV-Vis Spectroscopic Study
Absorption spectroscopy is used to determine the optical properties of the s sized samples, which is an important factor for photocatalysts. Figure 2 show DR/UV-Vis diffuse reflectance spectra and the Kubelka-Munk plot [28] of the sa Ru/g-C3N4 and g-C3N4. The significant decrease in the reflection upon Ru doping ( 2a) can be attributed to the enhanced light absorption, which is in accordance w much darker color. Absorption in the visible light range (from 400 to 750 nm) is an important con for the photocatalytic activity of g-C3N4 under visible light. When doping Ru NPs on Absorption in the visible light range (from 400 to 750 nm) is an important condition for the photocatalytic activity of g-C 3 N 4 under visible light. When doping Ru NPs onto the material, an increased band gap (Eg) from 2.63 to 2.78 eV was determined (Table 1), corresponding to the blue shift of the absorption edge from 471 nm (g-C 3 N 4 ) to 446 nm Ru/g-C 3 N 4 ). However, both the g-C 3 N 4 and Ru/g-C 3 N 4 composites showed absorption in the visible region, corresponding to the transition from the valence band (VB) to the conduction band (CB). The E VB and E CB values were estimated by adopting the Mulliken electronegative principle, using Equations (1) and (2): where χ represents Mulliken electronegative symbol of g-C 3 N 4 (4.67 eV) and E e is the energy of free electrons on the hydrogen scale (E e ≈ 4.50 eV). In this calculation, for both g-C 3 N 4 and Ru/g-C 3 N 4 , the same electronegativity (χ) was taken because Ru doping (with 1.4% surface concentration) could just slightly modify it. This estimation resulted in an E CB value for this semiconductor which is more negative than that of g-C 3 N 4 . The obtained E CB values hardly differ, which results in an equilibrium of the electrons between the two CBs. Nevertheless, it can promote some capturing of the electrons from the CB of g-C 3 N 4 , which can be excited easier than Ru/g-C 3 N 4 . Thus, the charge recombination in the previous one becomes more hindered. The slightly more negative CB potential of the ruthenium-modified semiconductor favors the reduction of Cr(VI).

SEM and EDS Analysis
The morphology of g-C 3 N 4 consists of large sheet-like layers with folds and voids on the surface (Figure 3a). After adding the ruthenium (presumably RuO x ) nanoparticles, on the samples of Ru/g-C 3 N 4 composites appeared some nanoparticles agglomerated on the surface of g-C 3 N 4 , leading to the formation of a heteromorphic structure; in addition, the pore size was narrowed due to the covering of ruthenium on g-C 3 N 4 ( Figure 3b). material, an increased band gap (Eg) from 2.63 to 2.78 eV was determined (Table 1 responding to the blue shift of the absorption edge from 471 nm (g-C3N4) to 446 nm C3N4). However, both the g-C3N4 and Ru/g-C3N4 composites showed absorption visible region, corresponding to the transition from the valence band (VB) to the co tion band (CB). The EVB and ECB values were estimated by adopting the Mulliken electronegativ ciple, using Equations (1) and (2): where χ represents Mulliken electronegative symbol of g-C3N4 (4.67 eV) and Ee is t ergy of free electrons on the hydrogen scale (Ee ≈ 4.50 eV).
In this calculation, for both g-C3N4 and Ru/g-C3N4, the same electronegativity ( taken because Ru doping (with 1.4% surface concentration) could just slightly mod This estimation resulted in an ECB value for this semiconductor which is more ne than that of g-C3N4. The obtained ECB values hardly differ, which results in an equili of the electrons between the two CBs. Nevertheless, it can promote some capturing electrons from the CB of g-C3N4, which can be excited easier than Ru/g-C3N4. Thu charge recombination in the previous one becomes more hindered. The slightly mor ative CB potential of the ruthenium-modified semiconductor favors the reduct Cr(VI).

SEM and EDS Analysis
The morphology of g-C3N4 consists of large sheet-like layers with folds and vo the surface (Figure 3a). After adding the ruthenium (presumably RuOx) nanopartic the samples of Ru/g-C3N4 composites appeared some nanoparticles agglomerated surface of g-C3N4, leading to the formation of a heteromorphic structure; in additio pore size was narrowed due to the covering of ruthenium on g-C3N4 (Figure 3b).  In addition, the results of the EDS analysis ( Figure 4) also confirmed the prese ruthenium on the surface of g-C3N4 material. The elemental composition of the comp In addition, the results of the EDS analysis ( Figure 4) also confirmed the presence of ruthenium on the surface of g-C 3 N 4 material. The elemental composition of the compound Ru/g-C 3 N 4 synthesized from energy dispersive spectroscopy (EDS) shows that the compound contains 21.56% C, 3.21% O, 74.73% N, and 0.5% Ru. Ru/g-C3N4 synthesized from energy dispersive spectroscopy (EDS) shows that the compound contains 21.56% C, 3.21% O, 74.73% N, and 0.5% Ru.  Figure 5 is the FT-IR spectrum of samples g-C3N4 and Ru/g-C3N4. The characteristic absorption fringes at 3100 cm −1 on the IR spectra of g-C3N4 and Ru-g-C3N4, respectively, are assigned to the valence oscillations of N-H [29]. Absorption fringes at about 1637 cm −1 and 1242 cm −1 can be attributed to the fluctuations of the C-N, C=N valences of the aromatic heterocyclic [30]. The intense bands at 1637, 1572, 1409, and 1242 cm −1 were assigned to typical stretching vibration modes of triazine-derived repeating units Finally, the strong absorption fringes at 810 cm −1 in the g-C3N4 and Ru/g-C3N4 samples, respectively, characterize the oscillation of the s-triazine ring absorption band. Furthermore, there is no absorption pattern associated with sulfur bonds (such as -SH, -SN, -SC), demonstrating that elemental sulfur is completely liberated during heat treatment [31]. A similar mode vibration was also present in Ru/g-C3N4, clearly indicating all modes of vibration preservation after ruthenium incorporation and without disturbing the typical molecular structure of g-C3N4.  Figure 5 is the FT-IR spectrum of samples g-C 3 N 4 and Ru/g-C 3 N 4 . The characteristic absorption fringes at 3100 cm −1 on the IR spectra of g-C 3 N 4 and Ru-g-C 3 N 4 , respectively, are assigned to the valence oscillations of N-H [29]. Absorption fringes at about 1637 cm −1 and 1242 cm −1 can be attributed to the fluctuations of the C-N, C=N valences of the aromatic heterocyclic [30]. The intense bands at 1637, 1572, 1409, and 1242 cm −1 were assigned to typical stretching vibration modes of triazine-derived repeating units Finally, the strong absorption fringes at 810 cm −1 in the g-C 3 N 4 and Ru/g-C 3 N 4 samples, respectively, characterize the oscillation of the s-triazine ring absorption band. Furthermore, there is no absorption pattern associated with sulfur bonds (such as -SH, -SN, -SC), demonstrating that elemental sulfur is completely liberated during heat treatment [31]. A similar mode vibration was also present in Ru/g-C 3 N 4 , clearly indicating all modes of vibration preservation after ruthenium incorporation and without disturbing the typical molecular structure of g-C 3 N 4 .

XPS Analysis
XPS spectroscopy was employed to check the surface compositions as well as chemical state of the elements present in g-C 3 N 4 and Ru/g-C 3 N 4 .
Accordingly, C1s spectrum shows four characteristic peak components at 293.4, 288.2, 286.7, and 285.1 eV corresponding to plasmon excitation of the heptazine heterocycles, sp 2bonded carbon (N-C=N), C-O, and C-C bonds [32,33], respectively, in both the prepared g-C 3 N 4 and Ru/g-C 3 N 4 material (Figure 6a,b). It demonstrates that the structure of g-C 3 N 4 remains after synthesis.
The N1s spectra can be deconvoluted into four individual peaks at about 398.7, 400.0, 401.2, and 404.9 eV (Figure 6e). The peak at about 398.7 eV corresponds to the nitrogen atoms bound with three C atoms, N-(C) 3 , and the peak at 400.0 eV is attributed to C-N-C in the heptazine rings. The peak at 401.2 eV is assigned to the C-N-H bond. The peak at 404.9 eV is assigned to plasmon excitation of the aromatic system of heptazine heterocycles [32,33].

XPS Analysis
XPS spectroscopy was employed to check the surface compositions as well as ical state of the elements present in g-C3N4 and Ru/g-C3N4.
Accordingly, C1s spectrum shows four characteristic peak components at 288.2, 286.7, and 285.1 eV corresponding to plasmon excitation of the heptazine he cles, sp 2 -bonded carbon (N-C=N), C-O, and C-C bonds [32,33], respectively, in b prepared g-C3N4 and Ru/g-C3N4 material (Figure 6a,b). It demonstrates that the st of g-C3N4 remains after synthesis. The peaks at 530.5 eV and 531.7 eV in the O1s spectra are related to C-O and C=O groups, respectively ( Figure 6f). The binding energy at 533.3 eV is assigned to the adsorbed The surface atomic concentrations of the Ru/g-C 3 N 4 catalyst from the XPS analysis (Table S1 in the Supplementary Materials) show 1.4% ruthenium, which is almost three times higher than that determined by EDS. This deviation originates from the different penetrations of Ru during the impregnation, as well as from the measuring depths of the two techniques (3-10 nm for XPS, 3 µm for EDS). Due to the impregnation method, these concentrations are significantly lower than the theoretical 5%, in the case of which all Ru ions in the solution phase would have been deposited/incorporated.

TEM Analysis
The morphologies of representative material (g-C 3 N 4 and Ru/g-C 3 N 4 ) are determined by the TEM analysis technique. The distribution of ruthenium is shown in Figure 7. It is clearly visible that the morphology of the g-C 3 N 4 support (Figure 7a,b) prepared by the calcination method was of a lamellar structure. Figure 7c,d illustrate that the active catalyst component Ru exists in the form of nanoparticles (NPs) on the support and the particle size of Ru NPs. For the Ru/g-C3N4 catalyst, the C 1s/Ru 3d region has a complex structure indicating a clear peak at 281.8 eV (Figure 6b), which can be assigned-in conjunction with the Ru 3p3/2 peak at 463.3 eV (Figure 6d)-to the Ru(II)-nitrogen bond [34][35][36][37]. Ru/g-C 3 N 4 (f) Figure 6. XPS spectra: high resolution XPS spectra in the C1s region of g-C 3 N 4 (a) and of Ru/g-C 3 N 4 (b); full scan of Ru/g-(c) and Ru3p scan of Ru/g-C 3 N 4 (d); high resolution XPS spectra in the N1s region (e) and O1s region (f) of Ru/g-C 3 N 4 .

TEM Analysis
The morphologies of representative material (g-C3N4 and Ru/g-C3N4) are determined by the TEM analysis technique. The distribution of ruthenium is shown in Figure 7. It is clearly visible that the morphology of the g-C3N4 support (Figure 7a

Investigation of Adsorption Equilibrium Time
The absorbance change of the aqueous Cr(VI) over time due to adsorption onto Ru/g-C3N4 is shown in Figure 8. The adsorption equilibrium was achieved after 60 min. It was chosen as the suitable adsorption time for the subsequent photocatalysis experiments.

Investigation of Adsorption Equilibrium Time
The absorbance change of the aqueous Cr(VI) over time due to adsorption onto Ru/g-C 3 N 4 is shown in Figure 8. The adsorption equilibrium was achieved after 60 min. It was chosen as the suitable adsorption time for the subsequent photocatalysis experiments.

Investigation of Adsorption Equilibrium Time
The absorbance change of the aqueous Cr(VI) over time due to adsorption onto Ru/g-C3N4 is shown in Figure 8. The adsorption equilibrium was achieved after 60 min. It was chosen as the suitable adsorption time for the subsequent photocatalysis experiments.

Photocatalytic Activity of Ru/g-C3N4
Photocatalytic reduction experiments of Cr(VI) over g-C3N4 and Ru/g-C3N4 were performed under visible light irradiation (λ ≥ 400 nm) (Figure 9). The 5% Ru/g-C3N4 composite shows the highest photocatalytic activity, and 96.81% Cr(VI) is reduced after 120 min, which is two times higher than that of pure g-C3N4 (50.1%).

Photocatalytic Activity of Ru/g-C 3 N 4
Photocatalytic reduction experiments of Cr(VI) over g-C 3 N 4 and Ru/g-C 3 N 4 were performed under visible light irradiation (λ ≥ 400 nm) ( Figure 9). The 5% Ru/g-C 3 N 4 composite shows the highest photocatalytic activity, and 96.81% Cr(VI) is reduced after 120 min, which is two times higher than that of pure g-C 3 N 4 (50.1%). Table S2 (in the Supplementary Materials) displays a comparison on Cr(VI) removal by different photocatalytic methods. According to these data, Cr(VI) removal performance by photocatalysis based on Ru/g-C 3 N 4 is comparable to those of other photocatalytic materials [18,20,21,25,[38][39][40][41][42]. This proves that Ru doping into g-C3N4 gives positive results regarding the efficiency of the photocatalytic Cr(VI) reduction. This result may be attributed to the relation of the CB energies of g-C3N4 and Ru/g-C3N4 as indicated in Section 2.1.2. The very similar ECB values may lead to an equilibrium of the electrons between the two CBs. Hence, Ru/g-C3N4 has a good chance to entrap electrons from the CB of g-C3N4. (Notably, the latter one can be easier excited due to its lower band gap.) As a consequence of electron capturing, the chance for charge recombination is diminished. Additionally, the (even if slightly) more negative CB potential of Ru/g-C3N4 may promote a more efficient Cr(VI) reduction. Figure 10 shows a simplified scheme of the charge-transfer mechanism taking place in the system upon excitation of the g-C3N4 catalyst doped with Ru. On the basis of the XPS analysis, most of the ruthenium in +2 oxidation states are connected to N atoms through covalent bonds (see in Section 2.1.5.). This mechanism is also supported by the result of scavenging experiments (Section 2.3.4.). This proves that Ru doping into g-C 3 N 4 gives positive results regarding the efficiency of the photocatalytic Cr(VI) reduction. This result may be attributed to the relation of the CB energies of g-C 3 N 4 and Ru/g-C 3 N 4 as indicated in Section 2.1.2. The very similar E CB values may lead to an equilibrium of the electrons between the two CBs. Hence, Ru/g-C 3 N 4 has a good chance to entrap electrons from the CB of g-C 3 N 4 . (Notably, the latter one can be easier excited due to its lower band gap.) As a consequence of electron capturing, the chance for charge recombination is diminished. Additionally, the (even if slightly) more negative CB potential of Ru/g-C 3 N 4 may promote a more efficient Cr(VI) reduction. Figure 10 shows a simplified scheme of the charge-transfer mechanism taking place in the system upon excitation of the g-C 3 N 4 catalyst doped with Ru. On the basis of the XPS analysis, most of the ruthenium in +2 oxidation states are connected to N atoms through covalent bonds (see in Section 2.1.5.). This mechanism is also supported by the result of scavenging experiments (Section 2.3.4.).
Catalysts 2023, 13, x 11 of 20 Figure 10. Schematic illustration of the of the charge-transfer mechanism of the photocatalytic Cr(VI) reduction based on g-C3N4 c doped with ruthenium.

Effect of Initial Solution pH
As pH strongly affects the photocatalytic reduction of Cr(VI), the effect of pH was studied in the range 2-10 with Ru/g-C3N4. The optimal pH solution for effective Cr(VI) removal by Ru/g-C3N4 the material was 2 ( Figure 11 and Table 2).  In contrast, as the pH is increased, the Cr(VI) removal efficiency by Ru/g-C3N4 material is decreased. This can partly be explained by the surface charge property of the material which also depends on the pH. The pHPZC value of a material is the pH value at which

Effect of Initial Solution pH
As pH strongly affects the photocatalytic reduction of Cr(VI), the effect of pH was studied in the range 2-10 with Ru/g-C 3 N 4 . The optimal pH solution for effective Cr(VI) removal by Ru/g-C 3 N 4 the material was 2 ( Figure 11 and Table 2).

Effect of Initial Solution pH
As pH strongly affects the photocatalytic reduction of Cr(VI), the effect of pH was studied in the range 2-10 with Ru/g-C3N4. The optimal pH solution for effective Cr(VI) removal by Ru/g-C3N4 the material was 2 ( Figure 11 and Table 2).  In contrast, as the pH is increased, the Cr(VI) removal efficiency by Ru/g-C3N4 material is decreased. This can partly be explained by the surface charge property of the material which also depends on the pH. The pHPZC value of a material is the pH value at which  In contrast, as the pH is increased, the Cr(VI) removal efficiency by Ru/g-C 3 N 4 material is decreased. This can partly be explained by the surface charge property of the material which also depends on the pH. The pH PZC value of a material is the pH value at which the surface charge is zero. When the pH is lower than the pH PZC value, the catalyst surface becomes positively charged, resulting in better anion adsorption. Similarly, when the pH is higher than the pH PZC value, the surface of the adsorbent carries a negative charge, which will better adsorb the cations [33]. The pH PZC of Ru/g-C 3 N 4 was determined to be 6.59 (Figure 12). It is higher than the isoelectric point of the pristine g-C 3 N 4 (pH ZPC = 4-5 [44]), but the Ru 2+ species can promote the neutralization of the particle surface already at a lower proton concentration. At low pH, Cr(VI) forms exist as HCrO 4 − and Cr 2 O 7 2− ; when pH increases, they convert to CrO 4 2− . Furthermore, at low pH, the surface of the photocatalyst becomes highly protonated, leading to better adsorption of HCrO 4 − or Cr 2 O 7 2− . At higher pH, the surface of the photocatalyst becomes more negative, which tends to repel the negatively charged ions and thus decreases the photocatalytic reduction rate of Cr ( [45]. Hence, at high concentrations of H + ions, the oxidation potential of Cr(VI) dramatically increases compared to the cases of weakly acidic or neutral systems. Accordingly, the most significant increase in the efficiency of photocatalytic Cr(VI) reduction was observed at the pH change from 4 to 2 (see Figure 11). the surface charge is zero. When the pH is lower than the pHPZC value, the catalyst surface becomes positively charged, resulting in better anion adsorption. Similarly, when the pH is higher than the pHPZC value, the surface of the adsorbent carries a negative charge, which will better adsorb the cations [33]. The pHPZC of Ru/g-C3N4 was determined to be 6.59 (Figure 12). It is higher than the isoelectric point of the pristine g-C3N4 (pHZPC = 4-5 [44]), but the Ru 2+ species can promote the neutralization of the particle surface already at a lower proton concentration. At low pH, Cr(VI) forms exist as HCrO4 − and Cr2O7 2− ; when pH increases, they convert to CrO4 2− . Furthermore, at low pH, the surface of the photocatalyst becomes highly protonated, leading to better adsorption of HCrO4 − or Cr2O7 2− . At higher pH, the surface of the photocatalyst becomes more negative, which tends to repel the negatively charged ions and thus decreases the photocatalytic reduction rate of Cr(VI).  [45]. Hence, at high concentrations of H + ions, the oxidation potential of Cr(VI) dramatically increases compared to the cases of weakly acidic or neutral systems. Accordingly, the most significant increase in the efficiency of photocatalytic Cr(VI) reduction was observed at the pH change from 4 to 2 (see Figure 11).

Effect of Photocatalyst Concentration
The catalyst content is also one of the factors affecting the Cr(VI) treatment efficiency in water. Experiments were performed by varying the photocatalytic material content with a fixed Cr(VI) concentration (20 ppm) and an optimal pH was selected (pH 2). The amount of material selected is 1 g/L, 2 g/L, and 3 g/L, respectively. The results are shown in Figure 13.

Effect of Photocatalyst Concentration
The catalyst content is also one of the factors affecting the Cr(VI) treatment efficiency in water. Experiments were performed by varying the photocatalytic material content with a fixed Cr(VI) concentration (20 ppm) and an optimal pH was selected (pH 2). The amount of material selected is 1 g/L, 2 g/L, and 3 g/L, respectively. The results are shown in Figure 13.
When the material content was increased to 1 g/L, the treatment efficiency increased significantly. However, when increasing to 3 g/L, the processing efficiency tends to decrease. The influence of catalyst mass on the degradation process can be explained by the following reasons: an increase in catalyst mass leads to an increase in the number of active sites available on the catalyst surface, increasing the density of the catalyst particles in the illuminated area; therefore, the photocatalytic ability of the material is better, leading to a rapid increase in Cr(VI) treatment efficiency [33]. However, when the catalyst content increases, it leads to an increase in the density of particles suspended on the surface of the solution, hindering the penetration of light, and increasing the light scattering effect. In addition, when increasing the amount of catalyst added, each catalyst molecule has a reduced chance of contacting Cr 6+ because of the rapid reaction [46]. As a result, the efficiency and reaction rate can be improved with increasing catalyst content, but the Cr 6+ conversion capacity is reduced. When the material content was increased to 1 g/L, the treatment efficiency incr significantly. However, when increasing to 3 g/L, the processing efficiency tends crease. The influence of catalyst mass on the degradation process can be explained b following reasons: an increase in catalyst mass leads to an increase in the number of sites available on the catalyst surface, increasing the density of the catalyst particles illuminated area; therefore, the photocatalytic ability of the material is better, leadin rapid increase in Cr(VI) treatment efficiency [33]. However, when the catalyst conte creases, it leads to an increase in the density of particles suspended on the surface solution, hindering the penetration of light, and increasing the light scattering effe addition, when increasing the amount of catalyst added, each catalyst molecule ha duced chance of contacting Cr 6+ because of the rapid reaction [46]. As a result, th ciency and reaction rate can be improved with increasing catalyst content, but th conversion capacity is reduced.

Effect of Initial Cr(VI) Concentration
The photocatalytic efficiency of the Ru/g-C3N4 catalyst in the reduction of Cr(V investigated at various concentrations (15-100 ppm) of Cr(VI) at an initial solution 2.0, and 0.1 g/L of catalyst with a reaction time of 120 min ( Figure 14). The photocat efficiency of the reaction decreased when the initial concentration of Cr(VI) was incre and drastically reduced with 100 ppm of Cr(VI) (to ca. 10%). This behavior is relate large amount of Cr(VI) that is adsorbed on the surface of the photocatalyst and pre the light absorption during the reaction. Hence, under these conditions, the specifi available surface area of the photocatalyst decreased, which resulted in lower phot lytic efficiency. In this case, even the reaction rate (ppm/min) decreased signific However, at concentrations 30 and 50 ppm, the reaction rate increased compared case of 15 ppm because the faster capturing of CB electrons by Cr(VI) overcompen the decreased light absorption. From a comparison based on the product of the initia centration of Cr(VI) and the ln(C/C0) value at 120 min, the highest reaction rate w served at 50 ppm.

Effect of Initial Cr(VI) Concentration
The photocatalytic efficiency of the Ru/g-C 3 N 4 catalyst in the reduction of Cr(VI) was investigated at various concentrations (15-100 ppm) of Cr(VI) at an initial solution pH of 2.0, and 0.1 g/L of catalyst with a reaction time of 120 min ( Figure 14). The photocatalytic efficiency of the reaction decreased when the initial concentration of Cr(VI) was increased, and drastically reduced with 100 ppm of Cr(VI) (to ca. 10%). This behavior is related to a large amount of Cr(VI) that is adsorbed on the surface of the photocatalyst and prevents the light absorption during the reaction. Hence, under these conditions, the specific and available surface area of the photocatalyst decreased, which resulted in lower photocatalytic efficiency. In this case, even the reaction rate (ppm/min) decreased significantly. However, at concentrations 30 and 50 ppm, the reaction rate increased compared to the case of 15 ppm because the faster capturing of CB electrons by Cr(VI) overcompensated the decreased light absorption. From a comparison based on the product of the initial concentration of Cr(VI) and the ln(C/C 0 ) value at 120 min, the highest reaction rate was observed at 50 ppm.

Scavenging Effect of Coumarin
The mechanism of the photocatalytic Cr(VI) reduction involving a g-C3N4-based catalyst has earlier been investigated by application of various scavengers such as formic acid (for • OH radicals), p-benzoquinone (for superoxide radicals), silver nitrate (for electrons), and isopropanol (for holes) [20]. On the basis of the results of these free radical trapping

Scavenging Effect of Coumarin
The mechanism of the photocatalytic Cr(VI) reduction involving a g-C 3 N 4 -based catalyst has earlier been investigated by application of various scavengers such as formic acid (for • OH radicals), p-benzoquinone (for superoxide radicals), silver nitrate (for electrons), and isopropanol (for holes) [20]. On the basis of the results of these free radical trapping experiments, it was suggested that the generation of electrons and (in the presence of dissolved oxygen) O 2 •− plays a determining role in the primary steps of the photocatalytic reduction of Cr(VI) based on g-C 3 N 4 . Since, according to our previous studies, coumarin can efficiently react not only with • OH radicals [47] but with electrons, too [48,49], it was applied as a multifunctional scavenger competing for the primary agents photogenerated by excitation of Ru/g-C 3 N 4 . In the presence of coumarin with a concentration of 2.5 × 10 −5 M, the Cr(VI) treatment efficiency of Ru/g-C 3 N 4 decreased markedly (Figure 15). The performance reduced from 96.81% to 75.43%. These results indicate that from the two opposite processes (i.e., reaction with • OH radicals or even valence-bond holes, which would promote the reduction of Cr(VI), and capturing conduction-band electrons, which would decrease the efficiency of Cr(III) formation), the latter proved to be predominant. This observation is in accordance with the corresponding reaction rate constants of coumarin 6.4-6.9 × 10 9 mol −1 dm 3 s −1 with • OH radical and 1.1-1.7 × 10 10 mol −1 dm 3 s −1 with electron [50,51]). In addition, the shape of the curve regarding the Cr(VI) decay in the presence of coumarin (Figure 13a) indicates that competition for the conduction band electrons significantly hinders the Cr(VI) reduction, especially in the first hour of irradiation.
alyst has earlier been investigated by application of various scavengers such as formic (for • OH radicals), p-benzoquinone (for superoxide radicals), silver nitrate (for electr and isopropanol (for holes) [20]. On the basis of the results of these free radical trap experiments, it was suggested that the generation of electrons and (in the presence of solved oxygen) O2 • − plays a determining role in the primary steps of the photocata reduction of Cr(VI) based on g-C3N4. Since, according to our previous studies, coum can efficiently react not only with • OH radicals [47] but with electrons, too [48,49], it applied as a multifunctional scavenger competing for the primary agents photogener by excitation of Ru/g-C3N4. In the presence of coumarin with a concentration of 2.5 × M, the Cr(VI) treatment efficiency of Ru/g-C3N4 decreased markedly (Figure 15). The formance reduced from 96.81% to 75.43%. These results indicate that from the two o site processes (i.e., reaction with • OH radicals or even valence-bond holes, which w promote the reduction of Cr(VI), and capturing conduction-band electrons, which w decrease the efficiency of Cr(III) formation), the latter proved to be predominant. Thi servation is in accordance with the corresponding reaction rate constants of coumarin 6.9 × 10 9 mol −1 dm 3 s −1 with • OH radical and 1.1-1.7 × 10 10 mol −1 dm 3 s −1 with elec [50,51]). In addition, the shape of the curve regarding the Cr(VI) decay in the presen coumarin (Figure 13a) indicates that competition for the conduction band electrons nificantly hinders the Cr(VI) reduction, especially in the first hour of irradiation.

Reusability of the Photocatalyst
The reusability of the Ru/g-C 3 N 4 sample was evaluated regarding its stability over three consecutive cycles ( Figure 16). As Figure 16 indicates, the photocatalyst showed almost the same photodegradation activity towards 20 mg L −1 K 2 Cr 2 O 7 solution in all three experiments. A slight decrease in the removal efficiency may be due to a loss of catalyst during the collection after each run. Slight differences in the shape of the plot in the case second and third runs compared to the first one may probably arise from a temporary decrease of the active sites on the surface of the Ru/g-C 3 N 4 catalyst during its collection after each run. As a consequence, the adsorption-desorption property of the reused catalyst may slightly deviate from that of the fresh one. In the second period of the irradiations, however, a regeneration of the active sites took place; therefore, no significant decrease could be observed in the photocatalytic yield after three recycling experiments, indicating the stability of the catalyst. lyst may slightly deviate from that of the fresh one. In the second period of the irrad tions, however, a regeneration of the active sites took place; therefore, no significant crease could be observed in the photocatalytic yield after three recycling experiments, dicating the stability of the catalyst.

Preparation of g-C3N4 Catalyst
The g-C3N4 catalyst was synthesized by a simple calcination method: 3 g of thiou in a porcelain crucible with a cover. For the calcination process, the sample was heated to 550 °C at 2 °C min −1 for 4 h. At the end of the heating process, the crucible was coo down to room temperature, and the solid sample of g-C3N4 was ground to powder a collected [52]. Samples are stored in vials and labeled with sample g-C3N4. The sam synthesis process and the schematic for plausible intermediates at different temperat

Preparation of g-C 3 N 4 Catalyst
The g-C 3 N 4 catalyst was synthesized by a simple calcination method: 3 g of thiourea in a porcelain crucible with a cover. For the calcination process, the sample was heated 25 to 550 • C at 2 • C min −1 for 4 h. At the end of the heating process, the crucible was cooled down to room temperature, and the solid sample of g-C 3 N 4 was ground to powder and collected [52]. Samples are stored in vials and labeled with sample g-C 3 N 4. The sample synthesis process and the schematic for plausible intermediates at different temperature ranges have been shown in Figure 17. Notably, the g-C 3 N 4 can be synthesized by hydrothermal condensation from melamine or other triazine precursors, too. However, it was reported that the g-C 3 N 4 synthesized from thiourea, a sulfur-containing precursor, could take advantage of a sulfur-mediated process, which improved the degree of polycondensation and polymerization of g-C 3 N 4 , thus enhancing the energy conversion efficiency [53].

Preparation of Ru/g-C 3 N 4 Catalyst
The Ru/g-C 3 N 4 catalyst was synthesized by a simple method of ultrasonic impregnation. Initially, 0.95g g-C 3 N 4 was evenly dispersed in 34 mL H 2 O in a beaker, then 30 mL RuCl 3 aqueous solution with a concentration of 0.01 g mL −1 was added to the beaker. Subsequently, after ultrasound treatment for 60 min, 5 mL deionized water containing 300 mg NaBH 4 was added to the solution and kept the solution in ultrasound for another 120 min [52]. The solution was then filtered and washed several times with distilled water and alcohol to remove impurities. The material was dried at 110 • C for 16 h, finally obtaining a black material with the symbol 5% Ru/g-C 3 N 4 . ranges have been shown in Figure 17. Notably, the g-C3N4 can be synthesized by hydrothermal condensation from melamine or other triazine precursors, too. However, it was reported that the g-C3N4 synthesized from thiourea, a sulfur-containing precursor, could take advantage of a sulfur-mediated process, which improved the degree of polycondensation and polymerization of g-C3N4, thus enhancing the energy conversion efficiency [53].

Preparation of Ru/g-C3N4 Catalyst
The Ru/g-C3N4 catalyst was synthesized by a simple method of ultrasonic impregnation. Initially, 0.95g g-C3N4 was evenly dispersed in 34 mL H2O in a beaker, then 30 mL RuCl3 aqueous solution with a concentration of 0.01 g mL −1 was added to the beaker. Subsequently, after ultrasound treatment for 60 min, 5 mL deionized water containing 300 mg NaBH4 was added to the solution and kept the solution in ultrasound for another 120 min [52]. The solution was then filtered and washed several times with distilled water and alcohol to remove impurities. The material was dried at 110 °C for 16 h, finally obtaining a black material with the symbol 5% Ru/g-C3N4.
The elemental composition and chemical state analysis of sample surfaces have been carried out by XPS, on a Thermo Scientific ESCALAB Xi + instrument (assembled in Brno, Czech Republic). A monochromatized Al K-alpha source (1486.6 eV), with a 650 µm spot size was used. The pressure of the analysis chamber was lower than 10 −9 mbar before conducting the experiment. On each sample, a wide range of spectra were collected (at analyzer pass energy of 80 eV) for surveying the elemental composition. For quantitative and chemical state analysis, high-resolution spectra (at 40 eV pass energy) were recorded for
The elemental composition and chemical state analysis of sample surfaces have been carried out by XPS, on a Thermo Scientific ESCALAB Xi + instrument (assembled in Brno, Czech Republic). A monochromatized Al K-alpha source (1486.6 eV), with a 650 µm spot size was used. The pressure of the analysis chamber was lower than 10 −9 mbar before conducting the experiment. On each sample, a wide range of spectra were collected (at analyzer pass energy of 80 eV) for surveying the elemental composition. For quantitative and chemical state analysis, high-resolution spectra (at 40 eV pass energy) were recorded for the following photoelectron lines: C 1s, Ru 3d, N 1s, O 1s, Ru 3p, Ca 2p, and Na 1s regions. The charging of the sample surface was compensated for by using the automatic built-in charge compensation system. The energy of sp 2 -bonded C in N=C(-N) 2 , set at 288.2 eV was used as the internal reference for fine energy scale adjustment.
Samples for transmission electron microscopy (TEM) were prepared by depositing a drop of diluted aqueous suspension of the original samples on copper TEM grids covered by continuous carbon amorphous support film. TEM analyses were performed using a Talos F200X G2 instrument (Thermo Fisher Scientific, Waltham, MA, USA), operated at 200 kV accelerating voltage, equipped with a field emission gun and a four-detector Super-X energy-dispersive X-ray spectrometer (Termo Fisher Scientific), and capable of working in both conventional TEM and scanning transmission (STEM) modes. In our study, TEM bright-field images, HRTEM images, and STEM high-angle annular dark-field (HAADF) images were collected to visualize the crystal size, and the morphology of the particles, and HRTEM images as well as electron diffraction patterns were used to study the structural properties. STEM-EDS elemental maps were collected to measure and visualize the chemical compositions.

Photocatalytic Reduction Experiments
The photocatalytic activity of the as-prepared samples, g-C 3 N 4 and Ru/g-C 3 N 4 , was evaluated, which was assessed through the reductive efficiency of aqueous Cr(VI) to Cr(III) under the illumination with a 500W Hg-lamp through a cut-off filter transmitting λ ≥ 400 nm. The reaction was carried out at room temperature, in a glass double-shell reactor containing 50 mL of Cr(VI) solution (20.0 mg/L) and 0.1 g of photocatalyst. Prior to the light irradiation, the mixture was dispersed by sonication for 3 min and kept for stirring at 250 rpm for 60 min in the dark to attain the adsorption-desorption equilibrium. At regular intervals, 2 mL of the suspension was aspirated during irradiation and filtered with a syringe filter (0.45 µm nylon membrane) for immediate photocatalyst separation. The diphenylcarbazide was used as a reagent to analyze Cr(VI) concentration by UV-Vis spectrophotometry, observed at 540 nm by a spectrophotometer (Agilent 8453, Santa Clara, CA, USA) [54].

Conclusions
Ruthenium doping of g-C 3 N 4 prepared from thiourea proved to be a successful method for a significant enhancement of the efficiency of photocatalytic reduction of Cr(VI) based on this catalyst. According to the XPS analysis, most of the ruthenium incorporated by impregnation was mostly in oxidation state +2 and bound to the N atoms on or close to the surface of the catalyst particles. This modification can diminish the recombination of the charge carriers formed upon excitation of the photoactive semiconductor and, thus, increase the efficacy of the Cr(VI) reduction with CB electrons. pH 2 was found to be optimal for the photocatalytic reaction because of the advantageous surface charge and E(Cr(VI)/Cr(III)) redox potential. The performance of the Ru/g-C 3 N 4 catalyst was stable after several cycles of reuse, and its efficiency was comparable with those of other photocatalysts developed for Cr(VI) reduction. Hence, Ru-doped g-C 3 N 4 may be a promising photocatalyst for this purpose, and maybe for other reductive processes, too.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy.