Photocatalytic Hydrogen Production from Glycerol Aqueous Solution Using Cu-Doped ZnO under Visible Light Irradiation

Cu-doped ZnO photocatalysts at different Cu loadings were prepared by a precipitation method. The presence of Cu in the ZnO crystal lattice led to significant enhancement in photocatalytic activity for H2 production from an aqueous glycerol solution under visible light irradiation. The best Cu loading was found to be 1.08 mol %, which allowed achieving hydrogen production equal to 2600 μmol/L with an aqueous glycerol solution at 5 wt % initial concentration, the photocatalyst dosage equal to 1.5 g/L, and at the spontaneous pH of the solution (pH = 6). The hydrogen production rate was increased to about 4770 μmol/L by increasing the initial glycerol concentration up to 10 wt %. The obtained results evidenced that the optimized Cu-doped ZnO could be considered a suitable visible-light-active photocatalyst to be used in photocatalytic hydrogen production without the presence of noble metals in sample formulation.


Introduction
Hydrogen production from water under sunlight is, nowadays, one of the most ecological and rational alternatives for obtaining an energy carrier. Hydrogen is a zero-carbon-emission fuel and it is expected to be an important energy source in the near future [1]. Actually, hydrogen is mainly produced by natural gas through the steam methane reforming process [2]. Unfortunately, this approach involves the use of fossil fuels and CO 2 production, and for this reason, it is not considered sustainable. So, attention is increasingly being paid to eco-sustainable processes that can be carried out in mild conditions (at room temperature and pressure) and without the use of fossil fuels. In recent years, the photocatalytic process for hydrogen production from water has attracted considerable interest [3][4][5]. In particular, by means of heterogeneous photocatalysis, hydrogen can be produced mainly by two processes: the direct splitting of water into H 2 and O 2 [6,7] or the photoreforming of organic compounds [8]. This latter process represents a very attractive method for the removal of organic pollutants in wastewaters with the simultaneous valorization of these substances [9][10][11]. Photocatalytic H 2 production from water using organic substances as sacrificial agents has been studied since the 1980s [12]. For this purpose, different organic compounds such as methanol [13], ethanol [14], sugar [13], glycerol [15], and lactic acid [16] have been used. In particular, the use of glycerol as a sacrificial agent for H 2 production via the photoreforming process has aroused great interest [17,18]. The reason for this is the huge production of glycerol, the content of which in by-product streams from the biodiesel industry is about 10 wt %, while glycerol itself still has limited demand in the market [19]. In the literature, there are several papers on the efficiency of photocatalysis for hydrogen production starting from a glycerol aqueous solution [18,20,21]. In particular, different photocatalysts have been tested and the most investigated ones were TiO 2 -based materials [12,22]. However, the fast recombination of photogenerated electron-hole pairs and the large band gap of semiconductors may hinder their photocatalytic performances. Forming composites with another semiconductor or adding metal nanoparticles represents an interesting strategy that has been developed to solve this problem. For example, photocatalytic systems, such as TiO 2 -CuO x , TiO 2 -Pt, TiO 2 -Pd, and NiO-TiO 2 , have been studied for photocatalytic hydrogen production from a glycerol aqueous solution [21,23]. These synthesized photocatalysts proved to be much more active with respect TiO 2 alone, since the separation of the electron-hole pairs was improved. In addition, with the ZnO-ZnS/graphene composite, using a glycerol solution at 10 wt % initial concentration, the hydrogen production was equal to 289 µmol/L after 3 h of UV irradiation time [24]. Interesting results have also been reported for the Ag 2 O/TiO 2 photocatalyst that, in the presence of glycerol and after 4 h of UV irradiation time, was able to produce 2100 µmol/L [25]. However, in order to use solar energy, the development of visible-light-driven photocatalytic systems is highly required. For this purpose, the use of a noble metal, such as Au, Ag, and Pt, as a cocatalyst element or for photocatalyst doping has been reported [26][27][28]. Generally, the effect of the presence of the noble metal is the plasmon resonance absorption property in the visible light region [29]. Specifically, the noble metal's roles are the enhancement of visible light absorption of the photocatalyst and the separation of photogenerated charges in semiconductors, such as TiO 2 or ZnO. For example, Pt/TiO 2 samples, under visible light and in the presence of glycerol (50 wt % initial concentration) and methanol, have produced about 7500 µmol/L of H 2 after 5 h of irradiation time, while in presence of only TiO 2 , the hydrogen production was equal to only 500 µmol/L [30]. However, practical applications of photocatalytic hydrogen production based on a noble metal cocatalyst are restricted due to their high cost. So, as an alternative, it has been proposed to use different metals that are less expensive than noble metals but with interesting properties. This is the case with copper, which, as demonstrated in the literature [23,31,32], is able to enhance the photocatalytic properties of semiconductors such as TiO 2 and ZnO, promoting hydrogen production even in the presence of visible light. It has been reported that Cu nanoparticles, loaded on the TiO 2 surface by the photodeposition method, enhanced visible light absorption due to the plasmon resonance effect and also acted as cocatalysts to separate photogenerated charges in TiO 2 [33]. As a result, Cu/TiO 2 photocatalysts exhibited enhanced photocatalytic hydrogen production under visible light from a glycerol aqueous solution [33]. In this case, the maximum hydrogen production rate was equal to 0.24 mmol h −1 g −1 .
However, an excessive addition of Cu decreased the hydrogen production rate, suggesting that copper nanoparticles on the TiO 2 surfaces hinder photon absorption [33]. Moreover, the role of copper as a cocatalyst or a doping element for semiconducting materials has been clarified in a recent review paper [34]. In particular, the review highlighted several papers concerning different species of copper used as cocatalysts for titanium dioxide [34]. The author showed that, generally, copper is present on the semiconductor surface, which in turn is coupled to another semiconductor such as graphene and alumina [35]. ZnO has recently generated much interest within the scientific community. In fact, this semiconductor is a promising material due to its environmental stability, low cost compared with other metal oxides [31], and good photocatalytic property [36,37]. ZnO has a band gap energy similar to TiO 2 , being equal to 3.2 eV. As a consequence, it is active mainly in the presence of UV light [38]. In order to make it active under visible light irradiation, different strategies have been proposed in the literature for ZnO doping. Among these, the introduction of different types of metal dopant (e.g., Co, Mn, and Ni) into a ZnO semiconductor has been reported [39]. Different papers report the use of Cu-doped ZnO as an effective photocatalytic material [31,40]. However, this photocatalyst was studied mainly for the degradation of organic pollutants from water and wastewater [40][41][42][43]. In the recent years, the effect of Cu ions in ZnO nanorod arrays for photoelectrochemical water splitting under visible light was investigated [44]. In this work, the authors reported a considerable photocurrent density equal to 18 µA/cm 2 at 0.8 V during the water splitting reaction, which was about 11 times higher than that of undoped ZnO nanorod arrays. However, no data about the effective hydrogen production was reported. On the other hand, the use of Cu-doped ZnO as a photocatalyst for H 2 production from a glycerol aqueous solution has not been extensively investigated yet. For this reason, the aim of this Appl. Sci. 2019, 9, 2741 3 of 13 work was to evaluate the influence of operating conditions on photocatalytic hydrogen production from glycerol solutions in the presence of a visible-light-active Cu-doped ZnO photocatalyst, previously optimized towards the photocatalytic degradation of methylene blue and oxidation of arsenite into arsenate [37].

Photocatalyst Preparation
The photocatalyst synthesis procedure was reported in our previous work [37]. ZnO and Cu-doped ZnO were synthesized by a simple precipitation method. In detail, 5 g of zinc acetate-dihydrate ZnC 4 H 6 O 4 (Aldrich Italy, 99%) was dissolved in 50 mL of distilled water. Once the zinc acetate had completely dissolved in water, a solution of NaOH, obtained by dissolving 2 g of NaOH in 25 mL of distilled water, was slowly added to the acetate solution in order to obtain a precipitate. In the case of Cu-doped ZnO, a specific amount of copper acetate hydrate Cu(CH 3 COO) 2 was dissolved into the solution of ZnC 4 H 6 O 4 beforehand to induce the precipitation by NaOH addition. Finally, the generated precipitate was centrifuged, washed, and calcined at 600 • C for 2 h.
The Cu nominal loading is expressed as molar percentage and it was evaluated through the following relationship: where: nCu is the number of moles of Cu(CH 3 COO) 2 used in the synthesis; and nZn is the number of moles of Zn(CH 3 COO) 2 2H 2 O used in the synthesis.
All the synthesized photocatalysts are listed in Table 1. These samples were well characterized in our previous work [37]. In particular, XRD analysis revealed the formation of ZnO with a hexagonal wurtzite structure for all the prepared samples. Moreover, with respect to undoped ZnO, a slight shift of XRD peaks towards a higher angle was observed for the Cu-doped ZnO photocatalysts (see Supplementary Materials, Figure S1). This phenomenon is due to the narrowing of the crystal lattice of ZnO, because Cu 2+ , which has a cationic radius smaller than Zn 2+ , can easily replace Zn 2+ in the ZnO crystal lattice [45]. The morphology of the samples was evaluated by SEM analysis (see Supplementary Materials, Figure S2). From the SEM images, it was possible to note that the doping process did not change the overall morphology of the photocatalysts. In particular, both undoped ZnO and the 1.08Cu_ZnO photocatalyst were characterized by nonuniform macroaggregates.
The UV-Vis diffuse reflectance measurements evidenced that the Cu doping induced an improvement in the absorption of UV light and a decrease in the band gap value ( Table 1), confirming that the Cu-doped photocatalysts can be activated by visible light, as evinced by the results of photocatalytic activity both in the degradation of methylene blue dye and in the oxidation of arsenite to arsenate [37].

Photocatalytic Activity Tests
The photocatalytic experiments for hydrogen production from glycerol aqueous matrices were carried out in a photocatalytic Pyrex cylindrical reactor (I D = 1.25 cm) equipped with a N 2 distributor device (Q = 0.122 NL/min) to assure the absence of O 2 during the tests. In a typical photocatalytic test, 0.0525 g of photocatalyst was suspended in 35 mL of a glycerol aqueous solution at 5 wt % of glycerol concentration. To ensure the complete mixing of the suspension in the reactor, a peristaltic pump was used. The photoreactor was irradiated with a strip of visible LEDs with the wavelength emission in the range of 400-600 nm (nominal power: 10 W; light intensity: 32 mW/cm 2 ). The LED strip was positioned around the external surface of the reactor so the light source uniformly irradiated the reaction volume. The suspension was left in dark conditions for 2 h to reach the adsorption-desorption equilibrium of glycerol on the photocatalyst surface. The effect of catalyst dosage, initial glycerol concentration, solution pH, and incident light intensity was evaluated. Moreover, the stability of the optimized photocatalyst after four reuse cycles was also evaluated. The analysis of the H 2 in the gaseous phase coming from the photoreactor during the irradiation time was performed by using a continuous analyzer (ABB Advance Optima AO2020) equipped with a thermal conductivity detector (TCD).

Influence of Cu Content on the Hydrogen Production under Visible Light
The photocatalytic hydrogen production was evaluated for undoped ZnO and Cu-doped ZnO photocatalysts under visible light from a glycerol aqueous solution at 5 wt % initial concentration and with a catalyst dosage equal to 1.5 g/L. Figure 1 reports the results obtained after 4 h of visible light irradiation. It is worthwhile to note that the hydrogen production obtained with undoped ZnO (612 µmol/L) was comparable to that obtained by photolysis alone (592 µmol/L). This result means that the undoped ZnO was not active in the presence of visible light for the hydrogen production due to its large band gap energy [46]. On the other hand, all of the Cu-doped ZnO samples showed positive results towards hydrogen production under visible light irradiation. In detail, from the data presented in Figure 1, it is possible to observe that the hydrogen production increased with the increase of Cu content up to 1.08 mol %, but a further increase of the dopant level resulted in a decrease of photocatalytic hydrogen production.
In particular, the 1.08Cu_ZnO photocatalyst showed a hydrogen production equal to 4180 µmol/L. Possibly, the improved photocatalytic performances observed up to 1.08 mol % Cu content were due to the inhibition of the recombination rate of the electron-hole pairs [47]. However, the increase in Cu content beyond the optimal value induced a worsening of photocatalytic performances, indicating that an excess of dopant content may act as a recombination center of the electron gap pairs, as reported in previous works [37,48,49]. Regarding the comparison with the available literature data, it should be pointed out that in many cases, besides glycerol, in the solution there is the presence of a further sacrificial agent such as methanol or the addition of noble metals, which act as cocatalysts for hydrogen production [23]. A hydrogen production of 2600 µmol/L is reported starting from a glycerol solution with 6 wt % initial concentration and using a Pt/TiO 2 photocatalyst [15]. In particular, the hydrogen production reported in [15] was lower than that achieved with the 1.08Cu_ZnO catalyst, which showed good efficiency under visible light and in the absence of noble metals, allowing it to achieve a hydrogen production equal to twice that found in the literature. In particular, the 1.08Cu_ZnO photocatalyst showed a hydrogen production equal to 4180 μmol/L. Possibly, the improved photocatalytic performances observed up to 1.08 mol % Cu content were due to the inhibition of the recombination rate of the electron-hole pairs [47]. However, the increase in Cu content beyond the optimal value induced a worsening of photocatalytic performances, indicating that an excess of dopant content may act as a recombination center of the electron gap pairs, as reported in previous works [37,48,49]. Regarding the comparison with the available literature data, it should be pointed out that in many cases, besides glycerol, in the solution there is the presence of a further sacrificial agent such as methanol or the addition of noble metals, which act as cocatalysts for hydrogen production [23]. A hydrogen production of 2600 μmol/L is reported starting from a glycerol solution with 6 wt % initial concentration and using a Pt/TiO2 photocatalyst [15]. In particular, the hydrogen production reported in [15] was lower than that achieved with the 1.08Cu_ZnO catalyst, which showed good efficiency under visible light and in the absence of noble metals, allowing it to achieve a hydrogen production equal to twice that found in the literature.
It is well known that the parameter that defines the ability of a photocatalyst in hydrogen production is the electronic structure [50]. Considering that the copper content in the 1.08Cu_ZnO photocatalyst was very low (1.2 mol %), it is possible to assume its electronegativity value was equal to that of the undoped ZnO (5.94 eV) [46]. It is well known that the parameter that defines the ability of a photocatalyst in hydrogen production is the electronic structure [50]. Considering that the copper content in the 1.08Cu_ZnO photocatalyst was very low (1.2 mol %), it is possible to assume its electronegativity value was equal to that of the undoped ZnO (5.94 eV) [46]. From the band gap value (E g ) of 1.08Cu_ZnO (2.92 eV), it was possible to calculate the values of conduction band (E CB ) and valence band (E VB ) energy through the Mulliken relationship [51]: where X is the semiconductor electronegativity.
The obtained results are presented in Figure 2.
As it is possible to observe from the E CB (−0.05 eV vs. NHE) and E VB (2.86 eV vs. NHE) values, the electronic structure of the 1.08Cu_ZnO photocatalyst satisfied the thermodynamic requirements for the water splitting reactions, with respect to the potentials for water oxidation/reduction reactions, reported also by Yerga et al. [52]. In fact, the E VB value was equal to 2.89 eV versus NHE, which was more positive than the oxidation potential O 2 /H 2 O (+1.23 eV vs. NHE), and the E CB value was equal to −0.05 eV versus NHE, which was more negative than the reduction potential H + /H 2 (0 eV). Therefore, these values made the 1.08Cu_ZnO catalyst able to produce hydrogen in the presence of visible light both from thermodynamic and kinetic points of view thanks to the presence of the metal dopant that inhibited the recombination rate of the electron-hole pairs. Appl. Sci. 2019, 9, x FOR PEER REVIEW 6 of 14 As it is possible to observe from the ECB (−0.05 eV vs. NHE) and EVB (2.86 eV vs. NHE) values, the electronic structure of the 1.08Cu_ZnO photocatalyst satisfied the thermodynamic requirements for the water splitting reactions, with respect to the potentials for water oxidation/reduction reactions, reported also by Yerga et al. [52]. In fact, the EVB value was equal to 2.89 eV versus NHE, which was more positive than the oxidation potential O2/H2O (+1.23 eV vs. NHE), and the ECB value was equal to −0.05 eV versus NHE, which was more negative than the reduction potential H + /H2 (0 eV). Therefore, these values made the 1.08Cu_ZnO catalyst able to produce hydrogen in the presence of visible light both from thermodynamic and kinetic points of view thanks to the presence of the metal dopant that inhibited the recombination rate of the electron-hole pairs.

Influence of 1.08Cu_ZnO Catalyst Dosage in Photocatalytic Hydrogen Production
The optimization of the photocatalyst dosage was obtained by testing different amounts of 1.08Cu_ZnO in the range of 0.75-3 g/L (Figure 3).   Figure 3 displays the hydrogen production for the different 1.08Cu_ZnO dosages as a function of visible light irradiation time. As expected, very low H2 production was observed for the photolysis test (without the photocatalyst). In the opposite way, the photocatalytic H2 production from the glycerol solution was significantly enhanced in the presence of the 1.08Cu_ZnO photocatalyst in the glycerol aqueous solution. At a fixed irradiation time, there was an increase of hydrogen production up to a 1.08Cu_ZnO dosage of 1.5 g/L (4180 μmol/L of H2 after 4 h of visible light irradiation). Beyond this value of catalyst dosage, the photocatalytic performances worsened. In particular, at 3 g/L of catalyst dosage, the hydrogen production was 1670 μmol/L lower than that obtained with 0.75  Figure 3 displays the hydrogen production for the different 1.08Cu_ZnO dosages as a function of visible light irradiation time. As expected, very low H 2 production was observed for the photolysis test (without the photocatalyst). In the opposite way, the photocatalytic H 2 production from the glycerol Appl. Sci. 2019, 9, 2741 7 of 13 solution was significantly enhanced in the presence of the 1.08Cu_ZnO photocatalyst in the glycerol aqueous solution. At a fixed irradiation time, there was an increase of hydrogen production up to a 1.08Cu_ZnO dosage of 1.5 g/L (4180 µmol/L of H 2 after 4 h of visible light irradiation). Beyond this value of catalyst dosage, the photocatalytic performances worsened. In particular, at 3 g/L of catalyst dosage, the hydrogen production was 1670 µmol/L lower than that obtained with 0.75 g/L of catalyst dosage. The worsening of photocatalytic activity may be explained by the increased opacity of the aqueous solution, which made light penetration through the solution increasingly difficult [53]. Therefore, the optimal catalyst dosage was 1.5 g/L and it was used to evaluate the influence of the initial glycerol concentration in aqueous solution and the effect of pH of solution.  The results showed that hydrogen production increased with the increase of the glycerol concentration up to 10 wt % and then decreased for 20 and 40 wt % of glycerol initial concentration. Thus, the optimum glycerol concentration is to be considered equal to 10 wt %, with hydrogen production of 4776 μmol/L, as reported. The effect of the sacrificial agent initial concentration for photocatalytic hydrogen production has been extensively discussed in the literature. In particular, some studies have evidenced that this behavior follows a Langmuir-type isotherm [54][55][56], meaning that the photocatalytic hydrogen production rate is controlled by saturation of active centers by the adsorbed glycerol molecules [32]. However, in the present study, the interpretation of the data through a Langmuir-type isotherm does not adapt to the obtained experimental results; rather, the results show the existence of an optimum glycerol concentration. The worsening of the photocatalytic activity beyond the optimal value of the initial glycerol concentration can be attributed to the blockage of the adsorption of H3O+ on the active site's surface [32]. Moreover, the existence of a maximum value for hydrogen production as a function of initial glycerol concentration could be also attributed to the action of glycerol as a quenching agent for ions and radicals generated during the irradiation [18].

Effect of pH on Photocatalytic Hydrogen Production
The effect of initial pH on H2 photocatalytic production was evaluated in the range of 2-10 using the 1.08Cu_ZnO photocatalyst ( Figure 5) with a glycerol initial concentration equal to 10 wt %.
The highest H2 production was obtained at pH equal to 6 (H2 production equal to 4776 μmol/L). It was evident that the change in pH had in any case worsened the overall production of hydrogen,  The results showed that hydrogen production increased with the increase of the glycerol concentration up to 10 wt % and then decreased for 20 and 40 wt % of glycerol initial concentration. Thus, the optimum glycerol concentration is to be considered equal to 10 wt %, with hydrogen production of 4776 µmol/L, as reported. The effect of the sacrificial agent initial concentration for photocatalytic hydrogen production has been extensively discussed in the literature. In particular, some studies have evidenced that this behavior follows a Langmuir-type isotherm [54][55][56], meaning that the photocatalytic hydrogen production rate is controlled by saturation of active centers by the adsorbed glycerol molecules [32]. However, in the present study, the interpretation of the data through a Langmuir-type isotherm does not adapt to the obtained experimental results; rather, the results show the existence of an optimum glycerol concentration. The worsening of the photocatalytic activity beyond the optimal value of the initial glycerol concentration can be attributed to the blockage of the adsorption of H 3 O+ on the active site's surface [32]. Moreover, the existence of a maximum value for hydrogen production as a function of initial glycerol concentration could be also attributed to the action of glycerol as a quenching agent for ions and radicals generated during the irradiation [18].

Effect of pH on Photocatalytic Hydrogen Production
The effect of initial pH on H 2 photocatalytic production was evaluated in the range of 2-10 using the 1.08Cu_ZnO photocatalyst ( Figure 5) with a glycerol initial concentration equal to 10 wt %.
The highest H 2 production was obtained at pH equal to 6 (H 2 production equal to 4776 µmol/L). It was evident that the change in pH had in any case worsened the overall production of hydrogen, which in acidic conditions was equal to 2771 µmol/L and in basic conditions was only 1764 µmol/L. Therefore, the best result was obtained by operating at the spontaneous pH of the glycerol aqueous solution (pH = 6). This can be seen as an advantage, particularly from an economic point of view, since it is not necessary to use additional chemicals to change the pH in order to improve the production of hydrogen. The increase of hydrogen production observed with the increase of initial pH from 2 up to 6 was consistent with the literature concerning glycerol photoreforming [17]. The best hydrogen evolution rate could be related to the adsorption of glycerol on the photocatalyst surface [17,18], the charging behavior of the semiconductor surface, the size of the aggregates of the photocatalyst particles, as well as the positions of the valence and conduction band levels of the semiconductor with respect to those of the redox couples in solution [57].

Influence of Visible Light Intensity on Photocatalytic Hydrogen Production
The influence of visible light intensity on photocatalytic hydrogen production was studied with an initial glycerol concentration equal to 10 wt % at the spontaneous pH of the solution and with the 1.08Cu_ZnO catalyst dosage equal to 1.5 g/L. In particular, the incident light intensity increased from 8 to 32 mW/cm 2 . The obtained results are displayed in Figure 6.

Influence of Visible Light Intensity on Photocatalytic Hydrogen Production
The influence of visible light intensity on photocatalytic hydrogen production was studied with an initial glycerol concentration equal to 10 wt % at the spontaneous pH of the solution and with the 1.08Cu_ZnO catalyst dosage equal to 1.5 g/L. In particular, the incident light intensity increased from 8 to 32 mW/cm 2 . The obtained results are displayed in Figure 6.
As expected, at a fixed irradiation time, photocatalytic hydrogen production increased as LED light intensity increased due to the photogeneration of more electrons and holes [58]. These results are in agreement with the literature data that highlight the effect of light intensity on the performances of photocatalytic processes [59,60].

Influence of Visible Light Intensity on Photocatalytic Hydrogen Production
The influence of visible light intensity on photocatalytic hydrogen production was studied with an initial glycerol concentration equal to 10 wt % at the spontaneous pH of the solution and with the 1.08Cu_ZnO catalyst dosage equal to 1.5 g/L. In particular, the incident light intensity increased from 8 to 32 mW/cm 2 . The obtained results are displayed in Figure 6.

Recyclability Tests
Recyclability is one of the most important aspects to be considered in the formulation of a photocatalyst [59,61]. To confirm the recyclability of the 1.08Cu_ZnO sample, the photocatalytic tests for hydrogen production were repeated up to four cycles (Figure 7). At the end of each test, the catalyst was recovered by centrifugation of the solution, it was dried at room temperature for 48 h, and any regeneration step was carried out on the recovered catalyst.
Appl. Sci. 2019, 9, x FOR PEER REVIEW 10 of 14 As expected, at a fixed irradiation time, photocatalytic hydrogen production increased as LED light intensity increased due to the photogeneration of more electrons and holes [58]. These results are in agreement with the literature data that highlight the effect of light intensity on the performances of photocatalytic processes [59,60].

Recyclability Tests
Recyclability is one of the most important aspects to be considered in the formulation of a photocatalyst [59,61]. To confirm the recyclability of the 1.08Cu_ZnO sample, the photocatalytic tests for hydrogen production were repeated up to four cycles (Figure 7). At the end of each test, the catalyst was recovered by centrifugation of the solution, it was dried at room temperature for 48 h, and any regeneration step was carried out on the recovered catalyst. For all the reuse cycles, photocatalytic hydrogen production was substantially unchanged, being in the range of 4750-4775 μmol/L. These results evidenced the stability of the Cu-doped ZnO sample in photocatalytic hydrogen production from a glycerol aqueous solution under visible light and that no photocorrosion phenomena (typical of ZnO-based photocatalysts [62][63][64]) occurred in the used operating conditions.

Conclusions
Cu-doped ZnO-based photocatalysts, prepared by a precipitation method, were studied in For all the reuse cycles, photocatalytic hydrogen production was substantially unchanged, being in the range of 4750-4775 µmol/L. These results evidenced the stability of the Cu-doped ZnO sample in photocatalytic hydrogen production from a glycerol aqueous solution under visible light and that no photocorrosion phenomena (typical of ZnO-based photocatalysts [62][63][64]) occurred in the used operating conditions.

Conclusions
Cu-doped ZnO-based photocatalysts, prepared by a precipitation method, were studied in photocatalytic hydrogen production under visible light from a glycerol aqueous solution. The influence of the Cu content and operating conditions (initial glycerol concentration, photocatalyst dosage, light intensity, and initial pH of solution) were assessed. The highest hydrogen production was obtained at an optimum Cu content of 1.08 mol % with 10 wt % glycerol concentration, at the spontaneous pH of the solution (pH = 6), and with a photocatalyst dosage of 1.5 g/L. The electron transfer paths, induced by visible light irradiation, underlined that the 1.08Cu_ZnO catalyst is able to produce hydrogen in the presence of visible light both from thermodynamic and kinetic points of view since the doping with Cu inhibits the recombination rate of the photogenerated electron-hole pairs. Moreover, the optimized photocatalyst has proved to be active for several reuse cycles, maintaining the same hydrogen production and evidencing the absence of photocorrosion phenomena in the optimized operating conditions.