Facet-Specific Photocatalytic Activity Enhancement of Cu2O Polyhedra Functionalized with 4-Ethynylanaline Resulting from Band Structure Tuning.

Cu2O rhombic dodecahedra, octahedra, and cubes were densely modified with conjugated 4-ethynylaniline (4-EA) for facet-dependent photocatalytic activity examination. Infrared spectroscopy affirms bonding of the acetylenic group of 4-EA onto the surface copper atoms. The photocatalytically inactive Cu2O cubes showed surprisingly high activity toward methyl orange photodegradation after 4-EA modification, while the already active Cu2O rhombic dodecahedra and octahedra exhibited a photocatalytic activity enhancement. Electron, hole, and radical scavenger experiments prove that the photocatalytic charge transport processes have occurred in the functionalized Cu2O cubes. Electrochemical impedance spectroscopy also indicates reduced charge transfer resistance of the functionalized Cu2O crystals. A band diagram constructed from UV-vis spectral and Mott-Schottky measurements reveals significant band energy shifts in all Cu2O samples after decorating with 4-EA. From density functional theory (DFT) calculations, a new band has emerged slightly above the valence band maximum within the band gap of Cu2O, which has been found to originate from 4-EA through band-decomposed charge density analysis. The increased charge density localized on the 4-EA molecule and the smallest electron transition energy to reach the 4-EA-generated band are factors making {100}-bound Cu2O cubes photocatalytically active. Proper molecular decoration represents a powerful approach to improving the photocatalytic efficiency of semiconductors.

Introduction

Semiconductor nanocrystals have widely displayed facet-dependent photocatalytic activity, electrical conductivity, and optical properties.[1−6] For instance, {110}-bound Cu2O rhombic dodecahedra are much more photocatalytically active than {111}-bound octahedra, whereas {100}-bound cubes are inactive.[7,8] The intriguing photocatalytic inactivity of Cu2O cubes results from lack of charge carriers reaching the {100} surfaces, revealed by electron, hole, and radical scavenger experiments.[8] Density functional theory (DFT) calculations have suggested the presence of a thin surface layer consisting of a few lattice planes (∼1.5 nm or less) with somewhat different band structures for various Cu2O faces.[9] Thus, one can rationalize the observed photocatalytic facet effects by presenting different degrees of band bending within this surface layer to indicate facet-specific barriers to charge transport across a particular crystal face. The fact that Cu2O and Ag2O polyhedra possess the same crystal structure but exhibit an opposite facet-dependent photocatalytic activity trend also supports the use of surface layer-induced band structure tuning, rather than surface free energy (γ{110} > γ{111} > γ{100} for Cu2O; γ{100} > γ{110} > γ{111} for Ag2O) or surface atomic arrangement, as a more general approach to understand the photocatalytic facet behaviors.[10−13] Furthermore, DFT calculations on tunable numbers of different Si, Ge, and GaAs lattice planes have revealed that their metal-like and semiconducting faces are related to subtle variations in bond length, bond geometry, and frontier orbital electron distribution within the thin surface layer, which may give rise to differences in the band structure.[14−16]

The photocatalytic inactivity of Cu2O cubes was established because Cu2O nanocubes decorated with gold nanoparticles remained inactive, whereas Cu2O octahedra and rhombic dodecahedra recorded the expected photocatalytic enhancement due to more efficient charge carrier separation.[7] Beyond metal particle decoration, deposition of ZnO, ZnS, CdS, and Ag3PO4 particles on Cu2O cubes still did not show any photocatalytic activity, despite their favorable band energy alignments to facilitate charge carrier separation, whereas these semiconductor heterostructures yielded both photocatalytic enhancement and suppression on Cu2O octahedra and rhombic dodecahedra.[2,17−20] Different combinations of contacting lattice planes at the heterojunctions producing varying degrees of interfacial band bending are believed to cause such unexpected photocatalytic behaviors. The next type of interesting photocatalytic system is surface modification with conjugated molecules as a charge transport mediator. Previously, Co complexes have been anchored on TiO2 and CuInS2/ZnS particles via conjugated linkers for photoelectrochemical and photocatalytic hydrogen production, respectively.[21,22] TiO2 modified with arginine facilitates electron transfer for enhanced dye photodegradation.[23] Photoinduced intervalence charge transfer between ethynylferrocene bound on TiO2 particles has also been observed.[24] However, how semiconductor crystal facets affect photoinduced charge transport through a conjugated molecule remains unexplored. On the basis of results obtained from previous examples, Cu2O cubes should stay inactive even after surface molecular modification due to significant barrier height to charge carrier transport across the {100} faces, whereas photocatalytic behaviors cannot be safely predicted for octahedra and rhombic dodecahedra.

In this study, we have functionalized Cu2O cubes, octahedra, and rhombic dodecahedra with 4-ethynylaniline (4-EA) with its acetylenic carbon bonding to Cu atoms on Cu2O. Photodegradation of methyl orange in the solution was used to probe successful electron transfer from Cu2O to 4-EA for the production of radical species. Surprisingly, 4-EA-functionalized Cu2O cubes showed remarkably high photocatalytic activity. Electron, hole, and radical scavenger tests were performed to verify the photocatalytic results. Enhanced photocatalytic properties were also observed for the 4-EA-bound Cu2O rhombic dodecahedra and octahedra, but they differ in the extents of activity enhancement with respect to molecular concentration. Electrochemical impedance spectroscopy was carried out to establish favorable charge transfer of all Cu2O crystals after surface 4-EA modification. Diffuse reflectance spectra and Mott–Schottky plots were collected to construct band diagrams of pristine and 4-EA-modified Cu2O crystals to evaluate how 4-EA enables Cu2O cubes to become photocatalytically active. The central idea is that the ultrathin surface layer plus a monolayer of 4-EA constitutes a new surface layer, which can significantly change the surface band bending. Taking this idea further, DFT calculations were conducted to present band structures and density of states of the (100), (110), and (111) planes of Cu2O with and without 4-EA modification. The energy gap between valence band maximum and the emerged band from 4-EA is useful to understand the sudden appearance of photocatalytic activity from Cu2O cubes. Charge density localized on the 4-EA molecule anchored on the (100) plane also provides good insight for effective charge transfer by the molecule.

Results and Discussion

Cu2O rhombic dodecahedra, octahedra, and cubes were synthesized following our reported procedures by preparing an aqueous solution of sodium dodecyl sulfate (SDS), CuCl2, NaOH, and NH2OH·HCl to grow the crystals at room temperature.[25,26] Adjusting the reagent amounts gives the particle shape control. Figure S1 in the Supporting Information shows scanning electron microscopy (SEM) images of the synthesized Cu2O rhombic dodecahedra, octahedra, and cubes. The particles have high shape and size uniformity. Their size distribution histograms are provided in Figure S2. The rhombic dodecahedra have an average opposite face length of 230 nm. The average opposite corner length of octahedra is 330 nm. The average edge length of cubes is 233 nm. After calculations of total surface area and surface copper atoms in 10 mg of each sample, molar ratios of 1:100, 1:500, and 1:1000 of surface Cu atoms to 4-EA were used for molecular functionalization on the Cu2O particles, yielding different molar concentrations of 4-EA in 10 mL of ethanol with dispersed particles (Table S1). These large amounts of 4-EA were added to ensure complete coverage of 4-EA on the Cu2O particles, so electron transfer must be mediated by the conjugated molecule. Next, for a fair photocatalytic activity comparison, a fixed total particle surface area of 0.03 m2 was chosen. From the number of particles having this total surface area, 6.90, 5.72, and 6.99 mg of 4-EA-functionalized Cu2O rhombic dodecahedra, octahedra, and cubes were used for the photocatalysis experiments, respectively (see Table S2).

Decoration of Cu2O crystals with 4-EA molecules was confirmed by Fourier-transform infrared spectroscopy (FT-IR) characterization. Figure gives the FT-IR spectra of 4-EA and Cu2O rhombic dodecahedra, octahedra, and cubes functionalized with 4-EA at a 1:500 molar ratio of surface Cu atoms to 4-EA molecules. 4-Ethynylaniline shows a peak at ṽ = 3261 cm–1 from the stretching vibrations of the acetylenic hydrogen (C≡C—H). The acetylenic carbon–carbon (C≡C) bond stretching peak appears at ṽ = 2096 cm–1. The peak at ṽ = 1508 cm–1 comes from benzene ring stretching. The peaks at ṽ = 3489 and 3387 cm–1 with a shoulder band are attributed to primary amine stretching, whereas the amine bending vibration peak shows up at ṽ = 1616 cm–1.[27−29] After surface modification, the absence of the acetylenic hydrogen peak confirms the bonding of 4-EA onto Cu2O particle surfaces. The sp carbon–carbon bond stretching peak has shifted in position to ṽ = 2235 cm–1 for 4-EA-modified Cu2O octahedra and cubes.[27] Moreover, the peak due to C—H stretching vibrations of the aromatic ring shows up at ṽ = 2919 cm–1 for the functionalized Cu2O octahedra and cubes.[29] Strangely, this peak is not identifiable for rhombic dodecahedra. The peaks of primary amine stretching have shifted in position to ṽ = 3577 and 3324 cm–1, 3489 and 3286 cm–1, and 3450 and 3356 cm–1 for the surface-modified Cu2O cubes, octahedra, and rhombic dodecahedra, respectively.[29] Finally, the existence of peaks in the region 1616–1508 cm–1 from amine bending and benzene ring stretching modes also supports the presence of 4-EA on the Cu2O particle surfaces. FT-IR spectra of pristine Cu2O cubes, octahedra, and rhombic dodecahedra have been reported.[30] The broad band around 3400–3450 cm–1 disappears after 4-EA functionalization. The characteristic peak shifts for 4-EA bonded to different Cu2O surfaces are discussed later. In addition, thermal gravimetric analysis (TGA) and differential scanning calorimetry (DSC) of Cu2O cubes and 4-EA-modified Cu2O cubes were performed to estimate the amount of 4-EA anchored on Cu2O crystals (Figure S3). Instead of weight loss, weight gain due to conversion of Cu2O to CuO, made evident by the formation of black powder after the thermal analysis, was recorded for both samples. Thus, the amount of 4-EA molecules on Cu2O crystals cannot be determined this way.

Figure 1

FT-IR spectra of 4-ethynylaniline and surface-modified Cu2O cubes, octahedra, and rhombic dodecahedra.

After confirmation of particle functionalization, photodegradation experiments were carried out. Figures S4–S6 provide UV–vis absorption spectra of MO as a function of photoirradiation time for pristine and functionalized Cu2O rhombic dodecahedra, octahedra, and cubes. The photocatalysis results are summarized in Figure . Cu2O rhombic dodecahedra took 65 min to decompose 90% of methyl orange. With increasing concentrations of 4-EA introduced, nearly complete MO degradation occurred in 40 min for rhombic dodecahedra with a surface Cu atoms to 4-EA molar ratio of 1:100, 30 min for the 1:500 sample, and only 25 min for the 1:1000 sample. Photodegradation extent reached 90% for pristine octahedra after 130 min. Moderate activity enhancement was observed after 4-EA modification on Cu2O octahedra with 95% degradation taking 120 min for the 1:100 sample and 100–110 min for the 1:500 and 1:1000 samples. These results validate efficient photoexcited charge carrier transport through the conjugated 4-ethynylaniline. Remarkably, the photocatalytically inactive Cu2O cubes became highly active after bonding with 4-EA. All functionalized samples show similar activity completing the photodegradation reaction in just 45–50 min. This photocatalytic activity is far better than that of the functionalized octahedra and approaches that of the 1:100 sample for rhombic dodecahedra. This is a huge photocatalytic activity difference with important implications. Previously, it has been suggested that the slight photocatalytic activity of Cu2O cubes may be caused by some imperfect cubes, the presence of particles exposing other facets, and possible formation of CuO.[1] It appears that unintended bonding of particles with molecular species in the solution during crystal growth may also give rise to photocatalytic activity or cause substantial activity enhancement. On the other hand, this molecular effect also presents opportunities to tune the photocatalytic and other charge transfer properties of crystals. For example, the formation of semiconductor heterojunctions with a favorable band energy alignment to facilitate exciton separation is a widely employed strategy to improve photocatalytic efficiency. By contrast, molecular functionalization is a rather unexplored research direction, especially when crystal facets are also considered.

Figure 2

Plots of the extents of photodegradation of methyl orange vs time for Cu2O (a) rhombic dodecahedra, (b) octahedra, and (c) cubes with and without surface modification.

After the photocatalysis experiment, SEM characterization shows that the functionalized cubes and octahedra have maintained their morphologies (see Figure S7 for the cube case), so it is not a deterioration of the {100} faces making the Cu2O cubes become photocatalytically active. However, all rhombic dodecahedral samples can exhibit a higher degree of face etching (data not shown), recognizing that the as-synthesized rhombic dodecahedra are not all structurally perfect due to the acidic solution condition.[26]Figure S8 gives X-ray diffraction (XRD) patterns of the pristine and functionalized Cu2O crystals before and after the photocatalysis experiment. There are no changes in the XRD patterns. All reflection peaks are observed because the particles were not aligned to show a preferred orientation of deposition on the substrate. FT-IR spectra also support the retention of 4-EA molecules on Cu2O octahedra after the photocatalysis experiment (Figure S9).

Because the display of this exceptional photocatalytic activity in the 4-EA-modified Cu2O cubes is quite surprising, it is necessary to confirm this result with more evidence. The photogenerated electrons and holes will migrate to the crystal surfaces and react with water and/or adsorbed oxygen to produce radical species, which then photodegrade molecules. Electron, hole, and radical scavenger tests can reveal if these photocatalytic events take place.[8]Figure presents UV–vis absorption spectra of MO as a function of photoirradiation time for the aryl alkyne-modified Cu2O cubes in the presence of CrO3 as an electron scavenger and Na2C2O4 as a hole scavenger. Without introducing scavengers, it takes the surface-modified Cu2O cubes 50 min to fully degrade MO. When just 1 μmol of electron scavenger is added, photodegradation practically stops. Upon adding 1 μmol of hole scavenger, photodegradation extent drops significantly to only 30% after 50 min. Further increasing the Na2C2O4 amount to 3 μmol, the photocatalytic activity is mostly ceased. The scavenger tests indicate that photogenerated electrons and holes are effectively being captured, and excitons indeed reach the crystal surfaces. Electron paramagnetic resonance (EPR) characterization can further confirm the photocatalytic activity of 4-EA-modified Cu2O cubes. Figure S10 gives EPR spectra of photoirradiated DMPO and 4-EA-functionalized Cu2O cubes. The appearance of the classical quartet of DMPO–OH indicates significant production of hydroxyl radicals (•OH) under light illumination. EPR signals from superoxide anion radicals (O2•–) were also detected.[7,10] These experiments verify that 4-EA-modified Cu2O cubes are indeed photocatalytically active.

Figure 3

UV–vis absorption spectra of methyl orange vs irradiation time using 4-EA-modified Cu2O cubes as the photocatalyst in the presence of (a) 1 μmol of CrO3, (b) 3 μmol of CrO3, (c) 1 μmol of Na2C2O4, (d) 3 μmol of Na2C2O4, and (e) without any scavenger. (f) Summary of the electron and hole scavenger tests.

Electrochemical impedance spectroscopic (EIS) measurements were performed to further evaluate changes in charge transfer resistance before and after 4-EA modification. Figure shows the obtained Nyquist impedance plots. For a Nyquist plot, a large semicircle diameter means greater resistance to electron flow at the electrode–electrolyte interface (or charge transfer resistance, RCT). The corresponding charge transfer resistance values are provided in Table S3. All particle shapes display reduced semicircle diameter and thus charge transfer resistance after 4-EA functionalization. In particular, cubes and octahedra show a large lowering of RCT. The ultrasmall semicircle for the functionalized octahedra may be linked to the highest surface Cu atom density of the Cu2O {111} surface for molecular binding. Although the photocatalytic activity enhancement of Cu2O octahedra with 4-EA decoration is moderate, nonetheless, the EIS results support the appearance of photocatalytic activity of Cu2O cubes after surface modification with 4-EA.

Figure 4

Nyquist impedance plots for the pristine and 4-EA-modified Cu2O (a) rhombic dodecahedra, (b) octahedra, and (c) cubes. (d) Summary of the Nyquist impedance plots.

To understand how 4-ethynylanaline tunes the band energies of Cu2O crystals giving cubes photocatalytic activity, a comparison of valence band and conduction band positions of all samples should be useful. This can be done by deriving their band gap values from UV–vis absorption spectra and valence band energies of the p-type semiconductor Cu2O from Mott–Schottky plots. Figure S11 offers UV–vis spectra of the pristine and functionalized Cu2O cubes, octahedra, and rhombic dodecahedra. The corresponding Tauc plots are obtained, showing that band gaps of cubes and octahedra have decreased notably after 4-EA decoration. When applying some potential, positive and negative charge will accumulate at the interface of semiconductor and electrolyte solution and yield capacitance. The Mott–Schottky equation gives the relationship between the applied potential and capacitance. The flat band potential and the majority carrier density of the semiconductor can also be estimated by using the Mott–Schottky equation.[31]where C is the space-charge capacitance of the semiconductor, e is the electron charge (1.602 × 10–19 C), ε is the dielectric constant (7.60 for Cu2O), ε0 is the permittivity of the vacuum (8.85 × 10–14 F cm–1), NA is the acceptor density (hole density in p-type Cu2O), E is the applied potential, Efb is the flat band potential, kB is Boltzmann’s constant, and T is the absolute temperature.

Figure presents Mott–Schottky plots of the Cu2O samples. The negative slopes indicate that they are p-type semiconductors. The value NA of Cu2O can be obtained from the slope of the linear part of the curve in the Mott–Schottky plot, where NA = 2/eεε0 (slope). For example, the determined slope for Cu2O rhombic dodecahedra is 3.79 × 1010 F–2 cm4 V–1, so the value of NA is calculated to be 4.9 × 1020 cm–3. The valence band of the semiconductor can be calculated using the following equation.[31]where EV is the valence band of semiconductor, EF is the Fermi level, and NV is the effective state density in the valence band. Under the flat band condition, EF is equal to Efb. The intercept of the Mott–Schottky plot is Efb (see Figure ). At room temperature, kBT/e is 0.0257 V. NV can be obtained from the equation below.[31]where the effective mass m* of holes is 0.58m0 for Cu2O with m0 as the mass of the free electron (9.1 × 10–34 kg).[31,32]h is Planck’s constant. The value of NV for Cu2O is calculated to be 1.11 × 1019 cm–3. Using the above calculation, Ev of Cu2O rhombic dodecahedra is 0.61 V vs Ag/AgCl. The corresponding conduction band (Ec) is calculated using this equation.where Eg is the semiconductor band gap. The band gap values of different Cu2O crystals are obtained from the Tauc plots. The calculated Ev and Ec of various Cu2O crystals are available in Table S4. All the potentials are calibrated to the reversible hydrogen electrode (RHE) according to the following equation. The measured pH value of the electrolyte is approximately 6.3.

Figure 5

Mott–Schottky plots of (a–c) pristine and (d–f) 4-EA-modified Cu2O crystals. The slope of each linear line is provided.

From the calculated valence band and conduction band energies in RHE scale, a band diagram of the Cu2O crystals is presented in Figure . Some relevant potentials involving radical formation and consumption reactions are also indicated. For Cu2O cubes, it appears that their band positions shifted to the more negative potential side are linked to their photocatalytic inactivity. However, one should be aware that more similar band energies for these Cu2O particle shapes have also been reported.[18,19] After 4-EA modification, the band energies also become considerably more positive with improved photodegradation efficiencies. Compared to octahedra, cubes and rhombic dodecahedra have shown appreciably larger shifts of their valence band and conduction band positions to more positive energies after 4-EA modification, which match with the experimental observations. It may be this shift in band energies making functionalized Cu2O cubes become photocatalytically active. However, one should recognize that the band diagram cannot explain the photocatalytic inactivity of the pristine Cu2O cubes, as the band energies suggest that photocatalyzed radical formation should be possible.

Figure 6

Band positions of pristine and 4-EA-modified Cu2O crystals with their valence band energies determined from the Mott–Schottky plots and band gaps from Tauc plots. Potentials of relevant reactions involving radical species are also shown.

The inadequacy of the conventional band diagram to explain facet-dependent properties of semiconductors is again illustrated here. Thus, it is necessary to introduce the notion of an ultrathin surface layer with facet-specific band bending to account for the experimental observations. Using the band energies from Figure , Figure a is an adjusted band diagram showing large upward band bending for the {100} surface to signify significant energy barrier to charge transfer across this face, while charge transfer through the {111} and {110} faces is favorable.[9] The central idea here is that band structure of Cu2O is tuned after 4-EA termination, and the new surface constitutes the original thin layer plus the 4-EA monolayer. This view emphasizes the ability of the molecular surface modification to band structure tuning, particularly for cubes now presenting downward band bending enabling photoexcited electron migration through the conjugated 4-ethynylaniline to react with water and dissolved oxygen producing radicals. This view points to exciting possibilities that other surface molecular structures can also facilitate charge transfer. Figure b depicts the photocatalytic inactivity of Cu2O cubes because the photogenerated electrons and holes cannot reach the {100} faces, so they then recombine. After 4-EA functionalization, charge carriers are able to exit the {100} faces and migrate through the molecules to lead to MO photodegradation, as illustrated in Figure c.

Figure 7

(a) Adjusted band diagram of Cu2O crystals before and after surface modification. Drawings showing different photocatalytic responses for (b) pristine and (c) 4-EA-modified Cu2O cubes. The black arrows indicate the flow of photoexcited electrons.

The possibility that 4-EA modification changes the surface band structure of Cu2O was investigated through DFT calculations. As shown in Figure , DFT calculations were conducted to estimate band structures and corresponding density of states (DOS) of the Cu2O (100), (110), and (111) surface planes before and after modifying with 4-EA molecules. The band structures and DOS are already different for the pristine (100), (110), and (111) surface planes, resulting from the structural features on these crystallographic surfaces. Notably, some distribution of states exists above the Fermi level (green line). This phenomenon is consistent with previous DFT calculations of Cu2O surfaces.[33−35] In Figure b, a new band corresponding to an energy of 0.35 eV can be found above the valence band maximum (VBM) with a value of 0.28 eV in the case of Cu2O {100} surface modified by the 4-EA molecule. To determine the assignment of this new band, the band-decomposed charge density of this band was calculated to present the charge density distribution in real space, as seen in Figure c. Most of the charges are distributed over the 4-EA molecule. This shows that the emerged band results from the 4-EA molecule on the Cu2O {100} surface. The 4-EA-generated band also appears for Cu2O {110} and {111} surfaces (see Figure e,h). This provides evidence that 4-EA alters the band structure and the corresponding DOS of all Cu2O surfaces. However, compared with the Cu2O (100) and (111) cases, the charge density distribution of Cu2O (110) is apparently delocalized, in which the charge distribution not only is on the 4-EA molecule but also extends into Cu2O. Interestingly, the energy difference between the 4-EA band and VBM for the three Cu2O surfaces is in the order of ΔE(111) (0.18 eV) > ΔE(110) (0.14 eV) > ΔE(100) (0.07 eV). This relationship coincides well with the largest change in the photodegradation efficiency of Cu2O cubes after surface modification compared with those of Cu2O rhombic dodecahedra and octahedra. This likely means that the energy required to inject electrons to a higher energy state above the valence band is greatly reduced for the 4-EA-modified Cu2O cubes. Similarly, the new band can also facilitate charge transfer for rhombic dodecahedra and octahedra. Thus, the DFT results show that the emergence of a new band within the band gap contributed by 4-EA and facet-specific charge density distribution around the molecule and the crystal surface together alters the {100} faces of Cu2O sufficiently to make charge transport across this surface possible. Finally, returning to the characteristic infrared peak shifts of 4-EA bonded to different Cu2O surfaces, the cause for the phenomenon may be that the surface band structures and electron density and distribution around the molecule and the underlying Cu2O atoms have been tuned significantly. More examples are needed to see if this phenomenon can be used as a chemical probe for the crystal facet effects.

Figure 8

(a, d, g) Band structures and density of states for Cu2O {100}, {110}, and {111} surfaces and (b, e, h) surfaces modified by the 4-EA molecule. Fermi level is set at zero in all band structures and density of states and is tagged with the green line. The red solid line is the valence band maximum, while the red dot line indicates additional DOS contributed by 4-EA. (c, f, i) The corresponding band-decomposed charge densities of the 4-EA band on Cu2O surfaces.

Conclusion

Cu2O cubes, octahedra, and rhombic dodecahedra were densely decorated with 4-ethynylanaline through binding of the acetylenic carbons to surface copper atoms. FT-IR characterization verified this surface molecular functionalization. The inactive Cu2O cubes become highly photocatalytically active after binding to 4-EA. 4-EA-modified Cu2O octahedra and rhombic dodecahedra also showed photocatalytic enhancement due to photoexcited charge transfer through the conjugated molecules. Electron, hole, and radical scavenger tests were performed to confirm the presence of photocatalytic activity of the functionalized cubes. Electrochemical impedance spectral measurements revealed a notable reduction in the charge transfer resistance for all the surface-modified Cu2O samples. Mott–Schottky plots and Tauc plots were obtained to construct band diagrams of the pristine and 4-EA-decorated Cu2O crystals. Significant band energy shifts are observed, which may contribute to the photocatalytic activity enhancement of Cu2O rhombic dodecahedra and sudden activity appearance of Cu2O cubes. Nevertheless, it is still necessary to use surface band bending to describe the photocatalytic activity properties of the functionalized Cu2O crystals, treating the crystal surface now as composed of the thin surface layer plus the 4-EA monolayer. The idea that the surface is altered with 4-EA decoration was supported by DFT calculations, showing the emergence of a new band located just 0.07–0.18 eV above the valence band maximum. Band-decomposed charge density analysis indicates that the new band is derived from 4-EA. The localization of charge density over the 4-EA molecule for the Cu2O {100} surface and the smallest band energy difference between the valence band and the 4-EA-generated band together tunes the cube surface considerably to make photocatalytic activity possible. It is envisioned that molecular functionalization can be a powerful design strategy to boost photocatalytic efficiency beyond metal cocatalyst and semiconductor heterojunction approaches.

Experimental Section

4-EA-Functionalized Cu2O Crystals

Cu2O cubes, octahedra, and rhombic dodecahedra were synthesized in aqueous solution following reported procedures (see the Supporting Information).[25,26] Using the average edge length for cubes, opposite corner length for octahedra, and opposite face length for rhombic dodecahedra, their surface areas and volumes were calculated (Figure S12). From the density of Cu2O (6.0 g/cm3), the number of particles and total surface copper atoms of each particle shape in 10 mg were obtained, using the surface copper atom densities reported before (10.98, 14.27, and 7.76 Cu atoms/nm2 for the (100), (111), and (110) planes of Cu2O, respectively).[30] Depending on the molar ratios of surface Cu atoms to 4-EA molecules, different weights of 4-EA were used to yield its molar concentrations in 10 mL of solution (see Table S1). Next, 10 mg of Cu2O particles were dispersed in 2.0 mL of 99% ethanol, followed by the addition of calculated amounts of 4-EA in 8 mL of ethanol. After stirring for 3 h at room temperature, the particles were washed with ethanol several times. After centrifugation at 7000 rpm for 5 min, particles were dried by purging with nitrogen before storage. No unexpected or unusually high safety hazards were encountered in the experiment.

Photocatalytic Activity Measurement

Surface-functionalized Cu2O particles having the same total surface area of 0.03 m2 were used for the comparison of their photocatalytic activities. Calculations give the particle weight for each shape (Table S2). Here, 6.99, 5.72, and 6.90 mg of Cu2O cubes, octahedra, and rhombic dodecahedra are needed, respectively. After pristine or functionalized Cu2O particles were dispersed in 40.5 mL of deionized water in a quartz cell, 4.5 mL of 150 ppm methyl orange (MO) solution was added into the cell, giving the MO concentration at 15 ppm. The quartz cell was illuminated by a 500 W xenon lamp with a UV-blocking filter placed 30 cm away from the cell. The light intensity reaching the cell was measured to be 300 mW/cm2 by a power meter. Photocatalytic decomposition of MO solution was traced by UV–vis absorption spectra. Each 1 mL of solution withdrawn from the cell in a fixed time interval was centrifuged at 7000 rpm for 3 min to remove the Cu2O particles. The upper liquid was monitored using UV–vis spectroscopy.

Electrochemical Measurements

The electrochemical impedance spectroscopy (EIS) and Mott–Schottky plots were measured by using a Zehner Zennium electrochemical workstation. The electrochemical cell consists of photocatalyst (nearly 2 mg) coated on an indium tin oxide (ITO) glass as the working electrode, Pt wire as the counter electrode, Ag/AgCl as the reference electrode, and 0.1 M Na2SO4 as the electrolytic solution. Cu2O particles covering an area of 1 × 2 cm2 were coated on the ITO glass, and silver ink was applied to the glass for electrical connection. After drying the silver ink, epoxy glue was applied over the ink for protection from the electrolytic solution. Electrochemical impedance spectroscopic measurements were carried out with a frequency range from 10 mHz to 100 kHz and an amplitude of 5 mV. Mott–Schottky plots were obtained at a frequency of 1 kHz and an amplitude of 5 mV to determine the flat band potential.

Instrumentation

SEM images of the samples were obtained by a JEOL JSM-7000F scanning electron microscope. XRD patterns were recorded on a Shimadzu XRD-6000 diffractometer with Cu Kα radiation. UV–vis absorption spectra were recorded by a JASCO V-670 spectrophotometer. An X500 xenon lamp from Blue Sky Technologies was used as an irradiation source in the photocatalytic experiments. EPR spectra were obtained on a Bruker ELEXSYSE 580 CW/Pulse spectrometer. FT-IR spectra were recorded on a Bruker Vertex 80/80v spectrometer. A Linseis Pt-1600 thermal analyzer was used to collect TGA/DSC data.

Density Functional Theory Calculation

Ab initio total-energy calculations within density functional theory were performed using the Vienna ab initio simulation package (VASP).[36,37] The ultrasoft pseudopotentials with the projector augmented wave method and the generalized-gradient approximations were used in the structure relaxation, electronic structure calculation, and band-decomposed charge density analysis. For calculations of the band structure and the corresponding density of states, we conducted the Perdew–Burke–Ernzerhof (PBE) functional to obtain the variation tendency of valence band maximum for band gap estimation.[38] The geometric structures of Cu2O and the 4-EA molecule were relaxed by conjugate gradient method to achieve their optimized lattice constants and bond lengths at the equilibrium state. The Brillouin zone was sampled with 12 × 12 × 12 Monkhorst–Pack k-point grids for the Cu2O crystal and 2 × 2 × 2 for 4-EA. The energy cutoff for plane waves was 680 eV for the Cu2O crystal and 4-EA molecule, and convergence criteria for the electronic and ionic relaxation were 10–5 and 10–4 eV, respectively.

Subsequently, {100}, {110}, and {111} surfaces of Cu2O crystals with and without 4-EA modification were created using the optimized supercell with a vacuum layer larger than 10 Å separating the periodic boundaries at the z direction. For surfaces modified by 4-EA molecules, the spacing of molecules is larger than 8–10 Å. The spacing can be adjusted to control the molecular coverage on the crystal surface. These Cu2O surface models were subjected to energy minimization calculations again with the k-point mesh of 3 × 3 × 1 and 400 eV energy cutoff in order to obtain the optimized geometric structures. In the calculations of band structure and corresponding density of states, the valence band maximum is defined as the Fermi energy.[39] To compare the difference between Cu2O surface models with and without 4-EA modification, the band energy or the state energy minus the Fermi energy, i.e., E – EFermi, is plotted.