Hexavalent chromium Cr(VI), which has high carcinogenicity and teratogenicity, is a common heavy metal pollutant with high stability in the natural environment. It is released in the effluents of various industrial activities, including electroplating, leather tanning, metal finishing, dyeing, textile production, and steel fabrication. The removal of Cr(VI) is of great importance and has stimulated the development of effective technology for wastewater treatment.

While recent innovative technology includes adsorption, ion exchange, membrane, separation, bioremediation, and chemical precipitation, photocatalysis is considered an attractive, efficient and clean strategy to reduce toxic Cr(VI) species to the less harmful Cr(III) species. Liu et al. ^[1] reported that more than 99.0% of Cr(VI) was removed in 5 h over a-Fe₂O₃ under visible light irradiation.

As an important p-type semiconductor with a direct band gap of 2.17 eV that is suitable for sunlight utilization, Cu₂O nanocrystals with different morphology have been successfully prepared and were demonstrated as useful for gas sensing, solar energy conversion, and use in lithium ion batteries and the photoactivated splitting of water into H₂ and O₂. Kuo et al. ^[2] reported a facile synthesis of Cu₂O nanocrystals with a systematic shape evolution. Sun et al. ^[3] investigated the crystal facet dependent effect of polyhedral Cu₂O microcrystals that exposed different index facets on the photodegradation of methyl orange. Li et al. ^[4] synthesized flower-like Cu₂O architectures with a high surface area and large pore volume, which exhibited high and stable photocatalytic activity for the reduction of Cr(VI).

The present work focused on comparing the activity for reducing Cr(VI) ions under visible light over Cu₂O cubes, octahedrons, and rhombic dodecahedrons, respectively, dominated by {100}, {111}, and {110} facets.

All chemical reagents in this work were analytical grade and used without further purification. Cubic, octahedral, and rhombic dodecahedral Cu₂O nanocrystals were prepared by reported wet chemical methods ^{[5, 6]}.

Cu₂O cubes (denoted as Cub) without a capping ligand were synthesized by adding 5.0 mL of NaOH aqueous solution (2.0 mol/L) into 50 mL of CuCl₂ aqueous solution (0.01 mol/L) under stirring. Then 5.0 mL ascorbic acid aqueous solution (0.6 mol/L) was added and the system was stirred for 5 h at 55 °C.

Octahedral (OctO) and rhombic dodecahedral (RhdO) Cu₂O nanocrystals capped with oleic acid were obtained as follows. 40 mL of CuSO₄ solution (1 mmol) was mixed with 20 mL absolute ethanol and 2.5 mL oleic acid for octahedron (4 mL oleic acid for rhombic dodecahedron). After heating to 100 °C, 10 mL of NaOH was added in 5 min, followed by adding 30 mL aqueous solution containing 3.42 g D-(+)-glucose under constant stirring for 1 h. The products from the round bottom flask were collected by centrifugation and washed with distilled water and absolute ethanol several times, and finally dried at 60 °C.

To remove the capping ligand oleic acid on the surface of the octahedral and rhombic dodecahedral Cu₂O nanocrystals, the samples OctO and RhdO were treated by a controlled oxidation treatment ^{[7, 8]}. In a typical procedure, a mixed gas of propylene (8%), oxygen (4%), and argon (88%) with a total flow rate of 20 mL/min was fed to 0.2 g of Cu₂O loaded in a quartz tube (inner diameter of 15 mm). The temperature was programmed at a ramp rate of 5 °C/min to 215 °C and kept there for 30 min. The resulting octahedral and rhombic dodecahedral Cu₂O nanocrystals with clean surfaces were denoted as Oct and Rhd, respectively.

The crystal phase of the as-prepared samples was characterized by a X-ray diffractometer (XRD, PANalytical X"pert PRO) using Cu K_α radiation (λ = 0.15406 nm) in the range of 20° to 80°. Microstructural analysis was performed using field emission scanning electron microscopy (FESEM, Hitachi S-4800). The BET surface area was determined using a Micromertitics Tristar 3000. FT-IR spectra were recorded on a Nicolet IS 5 spectrophotometer.

In the photocatalytic experiment, 0.05 g of Cu₂O was suspended in 100 mL of Cr(VI) solution 40 mg/L. The pH value was adjusted to 5.0 by H₂SO₄ solution. After being kept in the dark for 50 min to attain adsorption equilibrium, the suspension with 7 vol% absolute methanol was irradiated with an 18 W LED white light lamp. A portion of the reaction solution was withdrawn every 5 to 10 min, which was centrifuged and analyzed for the Cr(VI) concentration by the 1,5-diphenylcarbazide method on a Tianjin UV-752B spectrophotometer.

Figure 1 shows the SEM images of the Cu₂O cubes, octahedrons, and rhombic dodecahedrons with the sizes of a few hundred nm. Octahedral OctO and Oct with clean surfaces with eight {111} facets have almost the same morphology with the edge size of 400 to 800 nm, except that the flat surfaces of OctO become slightly coarse after cleaning off the capping oleic acid by propylene oxidation at 215 °C. These are shown in Fig. 1(a) and (b). Huang’s group ^{[7, 8]} adopted this method for the first time to remove the oleic acid capped on the surface of Cu₂O nanocrystals. Using SEM and XRD characterization, they found that the treatment had no effect on the morphology and structure of the Cu₂O nanocrystals other than getting rid of the capping agent, which was deduced from their XPS and IR spectra. Figure 1(c) shows rhombic dodecahedral Cu₂O nanocrystals RhdO with 12 exposed {110} facets which had 300 to 600 nm edges. Its counterpart Rhd with clean surfaces also has a similar shape and size to RhdO (not shown here). The edge lengths of the Cu₂O cubes with six {100} facets were 200 to 400 nm, as shown in Fig. 1(d).

Fig. 1. SEM images of Cu2O crystals. (a) OctO, octahedral Cu2O prepared with oleic acid; (b) Oct, octahedral Cu2O with clean surfaces; (c) RhdO, rhombic dodecahedral Cu2O prepared with oleic acid; (d) Cub, cubic Cu2O.

Figure 2 shows the XRD patterns of the Cu₂O nanocrystals. All the diffraction peaks were indexed by the standard cubic structure of Cu₂O (JCPDS No. 05-0667). It is well known that Cu₂O is oxidized to CuO at temperatures higher than 150 °C. The fact that no peak other than those of Cu₂O was detected in the XRD patterns of Oct and Rhd indicated that the treatment at 215 °C for 0.5 h under the mixed C₃H₆ gas prevented the oxidation of Cu₂O into CuO.

Fig. 2. XRD patterns of Cu2O crystals with different morphology.

To examine the effect of the mild oxidation for removing the surface capping oleic acid, the as-prepared OctO and Oct were characterized by FT-IR. The results obtained are the same as reported by Huang’s group ^{[7, 8]}: no band around 2900 cm^-1 corresponding to the vibration of C-H appeared, showing that the oxidation treatment had removed the surface oleic acid. IR spectroscopy has long been utilized as a powerful tool to provide supplementary information on the nature of copper oxides: the absorption peak at 610 cm^-1 was attributed to the stretching of Cu(I)-O and the band at ~530 cm^-1to the vibration of Cu(II)-O ^{[9, 10]}. We further compared the absorption features of OctO and Oct in the range of 400-1000 cm^-1. This is shown in Fig. 3. Both OctO and Oct displayed only one absorption peak at 605 cm^-1, which excluded the existence of CuO. Thus, IR characterization confirmed that OctO and Oct were of the cuprous oxide phase.

Fig. 3. FT-IR spectra of octahedral Cu2O prepared without (Oct) and with (OctO) oleic acid.

Figure 4(a) shows the catalytic performance for reducing Cr(VI) under visible light over the Cu₂O nanocrystals without a capping agent. The Cu₂O nanocrystals catalyzed the process, reducing 40% Cr(VI) in 3 h, as compared to that only 10% Cr(VI) was reduced under the same condition without the catalysts. From the data up to 100 min, the first order rate constant k for the Cu₂O samples with different morphology were obtained and shown in Fig. 5. The catalytic activity for the reduction of Cr(VI) was ranked as: rhombic dodecahedron > octahedron > cube, i.e., facets {110} > {111} > {100}.

To evaluate the effect of the capping oleic acid on the performance of the catalysts, octahedral and rhombic dodecahedral Cu₂O nanocrystals before and after the mild oxidation treatment in the mixed C₃H₆ gas were used to reduce Cr(VI) under the same condition with visible light irradiation. OctO and Oct gave a similar kinetic behavior in 100 min, as shown in Fig. 4(b), with almost the same rate constant (Fig. 5). The situation of RhdO and Rhd was the same as the Cu₂O octahedrons. Hua et al. ^[11] reported a surface blocking effect exerted by the capping ligand on the reduction behavior of H₂ and CO over Cu₂O nanocrystals. They found that different capping ligands gave different reduction temperatures. Unlike their findings for the gas reaction, our experimental results that the OctO and Oct, and RhdO and Rhd samples have the same rate constant showed that there was no blocking effect by the capping oleic acid on the surfaces of octahedral and rhombic dodecahedral Cu₂O on their catalytic performance in the aqueous system. It was shown recently that the vertices and edges of Cu₂O nanocrystals experienced erosion due to slight dissolution after several weeks in water ^[12]. The capping oleic acid can act as a protecting agent to slow down the dissolution of Cu₂O in a weak acid solution or the oxidization by holes formed by light absorption ^[13]. Thus an apparent activity at a long reaction time that was higher than the actual one could be produced by these Cu₂O nanocrystals capped with oleic acid, as shown in Fig. 4(b).

Fig. 4. Photocatalytic reduction of Cr(VI) over Cu2O nanocrystals under visible light.

The crystal structure, grain size, surface area, and morphology of the catalyst will affect the photocatalytic activity. In order to explore the factors determining the performance of the samples, the use of the photocatalytic activity per unit surface area, the specific activity, is more reasonable. The specific surface areas S_BET of cubic, octahedral, and rhombic dodecahedral Cu₂O determined by the BET method were 2.85, 0.65, and 0.96 m²/g, respectively. Their specific rate constants k’, defined as k/S_BET, for the photocatalytic reduction of Cr(VI) were calculated and shown in Fig. 5. The results clearly showed that the specific activity of the Cu₂O nanocrystals depended on the exposed crystal planes and can be ranked as {111} > {110} > {100}.

Fig. 5. First order rate constants for the photocatalytic reduction of Cr(VI) over Cu2O with different morphology.

Recent theoretical investigations have shown that there are differences between the different facets of Cu₂O crystals. While the {111} crystal plane of Cu₂O octahedrons has both coordination unsaturated Cu and coordination saturated Cu ^{[14, 15]}, the {110} crystal plane of Cu₂O rhombic dodecahedrons only has coordination saturated Cu and the {100} crystal plane of Cu₂O cubes has coordination unsaturated O. Therefore, the Cu₂O {111} and {110} facets would have a higher energy compared to the {100} plane with 100% terminal oxygen. As a result, the {111} and {110} facets with the higher surface energy would be expected to be more easily photoexcited to produce electrons and holes and have a higher photocatalytic activity than the {100} facets ^[16]. As to comparing the {111} and {110} facets of Cu₂O, it would be reasonable that the former with unsaturated Cu gives a better catalytic result because coordination unsaturated sites usually act as the active sites, which would benefit photo-generated electrons and holes migration to the surface and reduce the electron-hole recombination probability. More work is needed to elucidate the factors determining the facet dependent performance of metal oxide nanocrystals.

We systematically investigated the catalytic performance of Cu₂O cubes, octahedrons, and rhombic dodecahedrons with well-defined shape for the catalytic reduction of Cr(VI) under visible light irradiation. There was no surface blocking effect exerted by the capping ligand on their reduction performance of Cr(VI) in aqueous solution. The photocatalytic specific activity over the Cu₂O nanocrystals dominated by different facets followed the order {111} > {110} > {100}.