Facet-Dependent Cuprous Oxide Nanocrystals Decorated with Graphene as Durable Photocatalysts under Visible Light

Three morphologies (octahedral, hierarchical and rhombic dodecahedral) of crystal Cu2O with different facets ({111}, {111}/{110}, and {110}) incorporating graphene sheets (denoted as o-Cu2O-G, h-Cu2O-G and r-Cu2O-G, respectively) have been fabricated by using simple solution-phase techniques. Among these photocatalysts, the r-Cu2O-G possesses the best photocatalytic performance of 98% removal efficiency of methyl orange (MO) with outstanding kinetics for 120 min of visible light irradiation. This enhancement is mainly due to the dangling “Cu” atoms in the highly active {110} facets, resulting in the increased adsorption of negatively charged MO. More importantly, the unique interfacial structures of Cu2O rhombic dodecahedra connected to graphene nanosheets can not only decrease the recombination of electron-hole pairs but also stabilize the crystal structure of Cu2O, as verified by a series of spectroscopic analyses (e.g., X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), scanning electron microscopy (SEM) and transmission electron microscopy (TEM)). The effective photocatalysts developed in this work could be applied to the efficient decolorization of negatively charged organic dyes by employing solar energy.


Introduction
Metal oxide semiconductors have been intensively investigated for photocatalytic degradation of organic pollutants for many years [1][2][3]. However, these photocatalysts, e.g., TiO 2 and ZnO (band gap ∼ = 3.2 eV), can exhibit superior photocatalysis properties in the ultraviolet (UV) light region, leading to their limited practical application in wastewater treatments due to the restricted use of solar energy [4][5][6][7][8]. Cuprous oxide (Cu 2 O), which is a p-type semiconductor with a direct band gap of ca. 2.17 eV, has been widely studied as an efficient photocatalyst [9][10][11][12][13] because of its abundance, low cost, environmental-friendliness and good visible-light response. Nonetheless, the photocatalytic activity of Cu 2 O is constrained by fast recombination of the electron/hole (e/h) pairs [14] and has low durability [15]. Therefore, many studies have been dedicated to enhancing visible light-active photocatalysis by enhancing the segregation of electron-hole pairs. For instance, incorporating ions into the semiconductor [16][17][18][19][20][21], sensitization with dyer and surface complex [22][23][24], and coupling two or more semiconductors [25,26]. Moreover, well-defined facets of Cu 2 O which exhibit unique crystallographic properties related to different atomic terminated arrangements have been demonstrated to make a great improvement to photocatalysis [27][28][29]. The {110} facets of Cu 2 O were found to have a superior photocatalytic activity toward the degradation of methyl orange [27]. In addition, the Cu 2 O octahedra crystals consisting of {111} facets showed higher photocatalytic performance as compared to truncated cubic crystals abundant in {100} facets which has been attributed to the lower surface energy density of {100} facets than that of {111} facets [30]. In comparison to the pure Cu 2 O octahedra with {111} surfaces and pure cubes with {100} surfaces, the combination of 26-facet and 18-facet polyhedra with main {110} was observed to have better adsorption and photocatalytic activities [31]. However, the aforementioned Cu 2 O crystals may suffer from the deterioration of their crystal structure during long-term operation [32].
To further increase the stability of Cu 2 O nanocrystals, one of the possible methods was carried out by the incorporation of carbonaceous materials onto Cu 2 O [33][34][35]. Graphene is a well-known two-dimensional (2D) carbon material [36], which has distinguishing physiochemical properties such as a theoretical surface area (~2965 m 2 g −1 ) [37], high intrinsic electron mobility (2 × 10 5 cm 2 V −1 s −1 ) [38], and exceptional chemical durability [39][40][41]. Owing to these unique properties, visible light-driven photocatalysts based on the synthesis of Cu 2 O-graphene composites [35,42,43] for fuel production and pollutant degradation have been proposed. However, limited studies have been reported in terms of different facets of Cu 2 O-graphene nanoheterostructures on their visible light-responsive activity and corresponding durability.
In this research, three different morphologies of Cu 2 O crystals with low-index facets ({111}, {111}/{110}, and {110}) decorated with graphene sheets were prepared via simple wet-chemical methods. These as-synthesized Cu 2 O crystals were thoroughly characterized by a variety of analytical spectroscopies and used as visible light-driven photocatalysts in the degradation of methyl orange (MO).

Photocatalyts Preparation
Typically, the synthesis of graphene oxide (GO) was carried out by referring to a modified route described earlier [44]. For preparation of octahedral Cu 2 O [27] and octahedral Cu 2 O-graphene, ca. 88 mL of deionized water or 1% GO solution was mixed with ca. 1 mL of copper(II) chloride solution and 0.87 g of sodium dodecyl sulfate (SDS) solids under vigorous stirring until the dissolution of the SDS powder. Then, 8.5 mL of 0.2 M NH 2 OH·HCl and 2.5 mL of 1.0 M NaOH solution were consecutively added to the resulting mixture. Lastly, the precipitate was treated by centrifugation and dried under a vacuum. The aforementioned photocatalysts prepared by using deionized water and 1% GO solution in the synthesis process were labelled as o-Cu 2 O and o-Cu 2 O-G, respectively.
In terms of the synthesis of hierarchical facets of Cu 2 O [30] and hierarchical facets of Cu 2 O-graphene, 48 mL of 1% GO solution was mixed well with 1 mL of copper(II) chloride solution. Afterward, 40 mL of ethylene glycol (EG) was introduced to the aforementioned mixture. Then, 24 mL of 0.1 M NH 2 OH·HCl was added to the above solution for 10 min. The resultant solution was moved to a Teflon-lined stainless steel autoclave, followed by ramping from room temperature to 180 • C for 1 h. The solid products were filtrated, washed and dried at ambient temperature. The aforementioned samples prepared without and with 1% GO solution in the synthesis process were denoted as h-Cu 2 O and h-Cu 2 O-G, respectively.
For the preparation of rhombic dodecahedral Cu 2 O [27] and rhombic dodecahedral Cu 2 O-graphene, ca. 34.6 mL of deionized water or 1% GO solution was heated to 34 • C in a water bath. Then, ca. 2.5 mL of copper(II) chloride solution (0.1 M) and 0.44 g of SDS solids were added into the above mixture with continuous stirring. While the SDS was dissolved, 2.5 mL of NaOH solution was introduced, followed by adding 12 mL of 0.1 M NH 2 OH·HCl to the resulting mixture. Finally, the precipitated products were separated by centrifuge, washed by using a water-ethanol mixture and dried. The aforementioned samples prepared by using deionized water and 1% GO solution in the synthesis process were denoted as r-Cu 2 O and r-Cu 2 O-G, respectively.

Characterizations of Photocatalyts
Powder X-ray diffraction (XRD) patterns of the samples were examined by using a PANalytical X'Pert PRO diffractometer with Cu Kα radiation (λ = 1.541 Å). The elemental compositions of samples were analysed by X-ray photoelectron spectroscopy (XPS) using a Kratos AXIS Ultra DLD spectrometer (Kratos Analytical Ltd., Stretford, Manchester, UK) and monochromated Al Kα X-ray source. The morphologies of the photocatlysts were observed through a scanning electron microscope (SEM, JEOL-7000F, JEOL Ltd., Akishima, Tokyo, Japan) with an accelerating voltage of 20 kV. In addition, to further study the detailed surface characteristics of the Cu 2 O nanocrystals, a high-resolution transmission electron microscopy (TEM, JEOL 2100F, JEOL Ltd., Akishima, Tokyo, Japan) was carried out. The UV-visible (UV-Vis) diffuse reflection spectra of the samples were collected on a UV-Vis spectrophotometer (Varian, Cary 100, Palo Alto, CA, USA).

Photocatalytic Degradation of Organic Pollutants
Photocatalytic degradation tests of MO were performed at 25 • C in a photoreactor (100 mL) described previously [35]. The test solution was prepared via mixing 5.0 mg of photocatalysts into 80 mL of MO solution (15 mg L −1 ). Prior to irradiation, the aqueous solution was kept stirring in the dark for 2 h to establish the adsorption equilibrium. The pH value of the solution was 5.6. Then, the photocatalytic reaction was performed by using a 300 W Xe lamp combined with a 420 nm cutoff filter as the light source. About 2 mL of aliquots was periodically withdrawn for every 30 min, centrifuged to separate solid samples, and the variations of concentration were measured by using a Hitachi UV-Visible (UV-Vis) spectroscope (Model U-2910).

Results and Discussion
As displayed in Figure 1 In the previous study [27], the intensity of the (220) peak to that of the (200) peak (I (220) /I (200) ) is able to be used to evaluate the degree of crystal structure. For instance, the index of I (220) /I (200) is nearly 0.79 in terms of rhombic dodecahedra Cu 2 O. As can be seen in Figure  samples were analysed by X-ray photoelectron spectroscopy (XPS) using a Kratos AXIS Ultra DLD spectrometer (Kratos Analytical Ltd., Stretford, Manchester, UK) and monochromated Al Kα X-ray source. The morphologies of the photocatlysts were observed through a scanning electron microscope (SEM, JEOL-7000F, JEOL Ltd., Akishima, Tokyo, Japan) with an accelerating voltage of 20 kV. In addition, to further study the detailed surface characteristics of the Cu2O nanocrystals, a high-resolution transmission electron microscopy (TEM, JEOL 2100F, JEOL Ltd., Akishima, Tokyo, Japan) was carried out. The UV-visible (UV-Vis) diffuse reflection spectra of the samples were collected on a UV-Vis spectrophotometer (Varian, Cary 100, Palo Alto, CA, USA).

Photocatalytic Degradation of Organic Pollutants
Photocatalytic degradation tests of MO were performed at 25 °C in a photoreactor (100 mL) described previously [35]. The test solution was prepared via mixing 5.0 mg of photocatalysts into 80 mL of MO solution (15 mg L −1 ). Prior to irradiation, the aqueous solution was kept stirring in the dark for 2 h to establish the adsorption equilibrium. The pH value of the solution was 5.6. Then, the photocatalytic reaction was performed by using a 300 W Xe lamp combined with a 420 nm cutoff filter as the light source. About 2 mL of aliquots was periodically withdrawn for every 30 min, centrifuged to separate solid samples, and the variations of concentration were measured by using a Hitachi UV-Visible (UV-Vis) spectroscope (Model U-2910).

Results and Discussion
As displayed in Figure  The sharp diffraction peaks indicate that the high crystallinity of Cu2O with different morphologies in all samples can be prepared by using wet-chemical methods. In the previous study [27], the intensity of the (220) peak to that of the (200) peak (I(220)/I(200)) is able to be used to evaluate the degree of crystal structure. For instance, the index of I(220)/I(200) is nearly 0.79 in terms of rhombic dodecahedra Cu2O. As can be seen in Figure 1, while incorporating graphene onto various crystals of Cu2O (o-Cu2O-G, h-Cu2O-G and r-Cu2O-G), the value of I(220)/I(200) is practically unchanged, indicating that no apparent perturbation of crystal structure is observed. This result implies that the introduction of graphene may cause little impact on the growth of nanocrystal Cu2O in the samples. The morphologies of the crystal Cu2O, o-Cu2O-G, h-Cu2O-G and r-Cu2O-G were investigated by using field emission scanning electron microscopy (FESEM). As can be seen in Figure 2a     The surface atomic compositions and interfacial electronic states of the samples can be investigated by XPS analysis, as displayed in Figure 4. The peaks at 284.6, 286.5 and 288.4 eV in the C1s spectra (Figure 4a) are assigned to non-oxygenated, epoxy/hydroxyl and carboxyl carbons, respectively [45,46]. The intensities observed for the peaks at 286.5 and 288.4 eV in the o-Cu2O-G, h-Cu2O-G and r-Cu2O-G samples are decreased slightly in comparison to the pure graphene oxide sheets, suggesting that the graphene sheets may maintain the reduced states even with the existence of Cu2O nanocrystals in the samples. The XPS of spectra of Cu2p (Figure 4b) indicate the photocatalysts have the principal and satellite features at ca. 934 eV and 944 eV which are attributed to Cu(II), while the features at ca. 932 and 952 eV are assigned to Cu(I) 2p3/2 and Cu(I) 2p1/2 peaks [9], respectively. In addition, the O1s (see Figure 4c) peaks at ca. 532.5 and 530.3 eV are assigned to C-O and Cu-O bindings, respectively. As a result, the contents of copper(I) for o-Cu2O-G and r-Cu2O-G composites are much higher, indicating that Cu2O nanoparticles may exist stably when dispersed on graphene nanosheets, which can maintain the photocatalytic performance during photodegradation process [47]. Moreover, it is hard to attain bulk information of the atoms by using XPS because a small number of atomic layers on the surface are identified. In addition, it is noteworthy that the diffraction features of copper oxide could hardly be found for all photocatalysts by using the XRD, which is possibly because of the identification of the crystal structure in the bulk phase (see Figure 1).  [9], respectively. In addition, the O1s (see Figure 4c) peaks at ca. 532.5 and 530.3 eV are assigned to C-O and Cu-O bindings, respectively. As a result, the contents of copper(I) for o-Cu 2 O-G and r-Cu 2 O-G composites are much higher, indicating that Cu 2 O nanoparticles may exist stably when dispersed on graphene nanosheets, which can maintain the photocatalytic performance during photodegradation process [47]. Moreover, it is hard to attain bulk information of the atoms by using XPS because a small number of atomic layers on the surface are identified. In addition, it is noteworthy that the diffraction features of copper oxide could hardly be found for all photocatalysts by using the XRD, which is possibly because of the identification of the crystal structure in the bulk phase (see Figure 1). .06 eV, respectively. This finding may be due to their different crystal sizes and distinct exposed facets in the photocatalysts. Combining the outcomes of the XRD, TEM, SEM, XPS and DR UV-Vis spectra, various morphologies of crystal Cu 2 O incorporated with graphene nanosheets may be able to take advantage of natural light for decontamination and also be recycled for reuse after photocatalytic reactions. crystals, all the photoactivities of prepared catalysts are highly enhanced, suggesting that the existence of graphene owing to the unique interface contact between Cu 2 O facets and graphene [48] can enhance the photocatalytic efficiency. Among them, ca. 98% photodegradation of MO can be observed for the r-Cu 2 O-G photocatalysts. It has been reported that the level of conduction band (CB) potential for Cu 2 O (−3.0 eV vs. vacuum) [48] was greater than the level of lowest unoccupied molecular orbital (LUMO) for MO (−3.3 eV vs. vacuum) [49]. As a result, the electron transfer from the excited MO to the Cu 2 O can barely happen, i.e., the MO sensitization during photocatalysis should hardly occur. Moreover, kinetic studies can be used to confirm the photocatalytic activities of various catalysts. As can be seen in Figure 6b, the photocatalytic reactions follow pseudo-first order kinetics and the corresponding data for rate constants (k) can be attained.  It is crucial to explore the durability and reusability in terms of practical applications of these photocatalysts. We take r-Cu 2 O and r-Cu 2 O-G as the example photocatalysts to test the stability of MO degradation. As can be seen in Figure 7a, an obvious decrease of photocatalytic performance (ca. 54.8%, i.e., from 80.0 to 25.2%) can be found for r-Cu 2 O after three consecutive tests. However, only ca. 10% of efficiency decline has been observed for r-Cu 2 O-G, which is probably due to the assistance of graphene. It is noteworthy that no obvious dissimilarity can be found in the fresh and used r-Cu 2 O-G photocatalysts (see XRD patterns in Figure 7b) in which Cu 2 O is still the dominant species. The slight decline of catalytic performance in the r-Cu 2 O-G photocatalysts may be due to the occurrence of intermediates [46] during photoreaction. It should be noted that the synthesized r-Cu 2 O-G photocatalysts can perform the photodegradation of MO with the highest kinetic rate (degradation efficiency = ca. 98% within 120 min) upon the presence of ultra-low content of samples (0.06 g L −1 ) using visible light illumination. Moreover, these r-Cu 2 O-G photocatalysts can be obtained via a facile and low-cost liquid-phase method. Above all, our r-Cu 2 O-G nanocomposites also have a remarkable enhancement of their long-term durability by the assistance of 1 wt % of graphene. In comparison to previously reported Cu 2 O-based photocatalysts, the synthesized r-Cu 2 O-G photocatalysts with dominant {110} facets of crystal Cu 2 O exhibit an excellent degradation kinetic in a low concentration of photocatalysts under visible-light irradiation that could be practically used to make the best use of daylight for the remediation of organic wastewater. As shown in Figure 8, according to the aforementioned results a probable mechanism for MO degradation by crystal Cu 2 O-graphene photocatalysts can be proposed. Upon visible-light illumination, the crystal Cu 2 O (e.g., rhombic dodecahedra Cu 2 O) is excited to produce electrons (in the conduction band (CB)) and holes (in the valence band (VB)). Because of the stronger contact between crystal facets and the graphene surface, the photogenerated electrons can be rapidly transported to the graphene nanosheets. As a consequence of the small difference between the Fermi potential of graphene and the reduction potential of O 2 [51,52] since the number of dangling Cu atoms on the {110} plane per unit surface area was nearly 1.5 times higher than that on the {111} plane, resulting in a more positively charged surface on {110} facets [28]. Therefore, the superoxide radical anions, hydrogen peroxide and h + may be the dominant species governing the photodegradation of MO under visible-light irradiation. It should be noted that the graphene sheets not only serve as acceptors of the photogenerated electrons from Cu2O but also as stabilizers to prevent the crystal Cu2O from structural destruction under irradiation over a long period.

Conclusions
In this study, three different crystal facets of Cu2O photocatalysts decorated with graphene sheets were synthesized by using simple solution-phase methods. The order in terms of photocatalytic MO degradation was observed to be as follows: rhombic dodecahedral Cu2O-graphene (r-Cu2O-G) > hierarchical Cu2O-graphene (h-Cu2O-G) > octahedral Cu2O-graphene (o-Cu2O-G), which was demonstrated to be related to the morphologies and crystal structure of facets (i.e., the {110} facets are most active toward MO degradation). In earlier reports regarding Cu2O-based photocatalysts,

Conclusions
In this study, three different crystal facets of Cu 2 O photocatalysts decorated with graphene sheets were synthesized by using simple solution-phase methods. The order in terms of photocatalytic MO degradation was observed to be as follows: rhombic dodecahedral Cu 2 O-graphene (r-Cu 2 O-G) > hierarchical Cu 2 O-graphene (h-Cu 2 O-G) > octahedral Cu 2 O-graphene (o-Cu 2 O-G), which was demonstrated to be related to the morphologies and crystal structure of facets (i.e., the {110} facets are most active toward MO degradation). In earlier reports regarding Cu 2 O-based photocatalysts, the r-Cu 2 O-G photocatalysts possessed high kinetics of photocatalytic degradation, i.e., by using an ultra-low content of samples (0.06 g L −1 ) to reach 98% of MO photodegradation under 120 min irradiation of visible light. More importantly, the cycling tests indicate that the resulting r-Cu 2 O-G composites show a surpassing durability compared to pure r-Cu 2 O nanocrystals. These significant enhancements are possibly because of the unique interfacial interaction of rhombic dodecahedra Cu 2 O (more positively charged {110} facets) with the graphene nanosheets, which could lead to the effective isolation of electron/hole pairs, stabilization of the crystal Cu 2 O, and an increase of MO adsorption. Consequently, the development of a cost-effective and facile method to prepare r-Cu 2 O-G composites with mostly {110} facets and graphene sheets, which exhibit superior photocatalytic performance (kinetics and stability), offers the potential for a promising application in the treatment of organic wastewater by utilizing natural sunlight.