Truncated Edge Cuprous Oxide Cube Architecture for Reduction of Nitrophenols

Truncated cubic Cu-oxide nanocrystals with edge lengths of ca. 600 nm were prepared employing the low-cost copper (II) chloride as the precursor. XRD revealed that the truncated cubic Cu-oxide nanocrystals are present in both the CuI and Cu0 state. XPS characterisation gives insight into the amount of each state present in the as-prepared, oxidised, reduced and recovered catalyst species. The catalytic activity of the truncated cubic Cu-oxide nanocrystals was tested for the reduction of nitrophenols using NaBH4. After one catalytic cycle of the reduction of 4-nitrophenol, the activity almost halved. The SEM images revealed that the recovered catalyst showed some disfigurement of the structure, and XPS confirmed the reduction of the CuI to metallic Cu0.

at 55 °C. The precipitate was filtered and washed with water. After drying in a vacuum oven, the product was obtained in 44% (138.5 mg) yield.

Reduction of nitrophenols using NaBH 4
5 ml Nitrophenols (200 ppm) was added to 70 ml water in a round bottom flask. 5 ml Freshly prepared NaBH 4 (0.1 M, 19 mg in 5 ml water) was added to the water mixture. The time was started when 10 mg of the truncated cubic Cu 2 O catalyst was added. A UV-Vis spectrum was collected at suitable time intervals.

Results and Discussion
The truncated cubic Cu-oxide nanocrystals were prepared by a colloidal synthesis procedure using ascorbic acid to direct structure growth. The hybrid morphological features of these Cu-oxide nanocrystals could be determined from Scanning Electron Microscopy (SEM) images, see Figure 1 (Left). The truncated cubic Cu-oxide is enclosed by six {100} facets on the basal planes of the cube, while twelve {110} facets enclose the edges of the cube and eight {111} facets are located on the corners (see Figure 1 (middle) for a cartoon illustrating this). These Cu-oxide nanocrystals are monodispersed and fairly uniform, with an average edge length of 600 nm (see the Supplementary Information (Fig. 1S) for the histogram and related data).
The crystallinity and phase purity of the truncated cubic Cuoxide nanocrystals were analysed by X-ray Diffraction (XRD). The XRD pattern of the Cu-oxide nanocrystals is present in Figure 1 (Right). The reflections at 29 25 which is attributed to the structure having the truncated system. The presence of the truncated cubic Cu-oxide nanocrystals in the Cu 2 O form was confirmed. However, a small amount of Cu 0 was also detected at 34.7°, 41.9°, 52.2° and 73.3° in correlation to published results. 26 The peak relating to the Cu 0 appeared as a small shoulder of the main peak resulting from the Cu 2 O form (see insert Figure 1 (Right)). This peak indicates that the truncated cubic Cu-oxide nanocrystals are not exclusively present in the Cu I state but that some Cu 0 is also present. No CuO (Cu II state) was detected since no peak was detected at ca. 32.5°, which is the position where the strong {110} peak appears. 27 X-ray photoelectron spectroscopy (XPS) analysis was conducted on the truncated cubic Cu-oxide nanocrystals sample to confirm this finding further.
XPS was used to determine the atomic composition and oxidation state of the copper and the ratio of the different atomic species present on the surface of the nanocrystals. The wide scan spectra (showing all the elements present) is given in the Supplementary Information ( Figure S2 (A)). These photoelectron lines were charge corrected by shifting the spectra so that the lowest binding energy of the simulated adventitious carbon C 1s photoelectron line is set at 284.8 eV. The peak maximum of the Cu 2p envelopes were found at ca. 932 and 952 eV for the Cu 2p 3/2 and Cu 2p 1/2 photoelectron lines, with small shake-up features at ca. 10-12 eV higher than the main photoelectron lines ( Figure S2 (B)). These shake-up features typically correspond to the presence of Cu II , either as CuO or Cu(OH) 2 . Due to the difficulty distinguishing between metallic Cu, Cu I (in Cu 2 O) and Cu II (in CuO or Cu(OH) 2 ), the curve fitting parameter proposed by Biesinger 28 was employed to identify the species present. The visual representation of the simulated curve-fitting parameters used (relative binding energy, relative ratio of Full Width at Half Maximum and relative atomic %) as per Cu species is presented in Figure 2 (A). Simultaneously, the Table with the numerical data can be found in reference no. 28. The simulated fits imposed on the XPS data of the as-prepared truncated cubic Cu-oxide nanocrystals using the fitting parameters of Biesinger are shown in Figure 2 (B), and the data are summarised in Table 1. This fit confirms the XRD data, where metallic Cu (most probably inside the particle) and Cu 2 O (in the Cu I state) covering the outside of the particle is present. Cu(OH) 2 (in the Cu II state) was also observed in the XPS. However, the XRD analysis did not detect the characteristic peaks of Cu(OH) 2 (strongest peak presents at ca. 31.2°). 29 The relative percentages of the different species present are 52.0% of Cu 2 O, 30.9% Cu 0 and 17.1% Cu(OH) 2 .
In literature, it has been described that Cu 2 O (Cu(I)) nanoparticles can be oxidised to CuO (Cu(II)) upon exposure to air. 30 This oxidation is reported to be size-dependent; the smaller the nanoparticles, the more reactive they are for oxidisation in Figure 1 The SEM images (A), a cartoon indicating the facets enclosing the particle (B) and the XRD pattern (C) of the as-prepared truncated cubic Cu-oxide nanocrystals. Table 1 Cu 2p 3/2 simulated curve-fitting parameters of the as-prepared truncated cubic Cu-oxide nanocrystals after treatment with NaBH 4 . The numerical peaks labels correlate with the simulated fitted peaks in Cu(OH)2 and CuO spectra shown in Figure 2.  ReseaRch aRticle E. Erasmus 108 S. Afr. J. Chem., 2021, 75, 106-110 https://journals.co.za/content/journal/chem/ air." It is also mentioned that bulk Cu 2 O is not stable in air at room temperature and can easily be oxidised. However, this stability was not observed for the truncated cubic Cu-oxide nanocrystals (average edge length 600 nm) as verified by the absence of Cu(II) in the XPS and XRD. The O 1s oxygen peak of the truncated cubic Cu-oxide nanocrystals was also fitted with the suggested simulation parameters of Biesinger, 28 see Figure  S3 in the Supplementary Information. The O 1s photoelectron line of the Cu 2 O and Cu(OH) 2 was found at 530.4 and 531.8 eV, respectively.

As prep
The UV-Visible spectra of the Cu-oxide nanocrystals suspended in water exhibited a broad absorption peak with two local maximum points located at 454 and 472 nm (see Figure 3, A). The latter compares very well with the reported wavelength of maximum absorption of cubic Cu-oxide suspended in ethanol, found at 470 nm. 31 The classical Tauc equation 32 was employed to determine the apparent bandgap energy, E g ', of the Cu-oxide nanocrystals (the term apparently is used since this is not a thin film but nanocrystals): where α is the absorption coefficient (α = ln(T/d), where T = transmission and d = thickness, average edge length was used in this case), K is a constant, E p is photon energy (E p = hc/λ, h = Planck's constant and c = speed of light), and E g is the bandgap energy. The extrapolated value (E p at α = 0 ) of the straight line to the X-axis of the graph of (αE p ) 2 versus E p represents the apparent bandgap energy, which was found to be E g ' = 2.07 eV (see Figure S4 in the Supplementary Information). Figure 3 (B) shows the ATR FTIR spectrum of truncated cubic Cu-oxide nanocrystals. The transmittance frequency peak at 1105 cm -1 is attributed vibrational band of the -O-H group (of Cu(OH) 2 ), in correlation with published results. 33 The presence of this peak further confirms that Cu(OH) 2 is present in the sample. The lower frequency peaks at 700 cm -1 and 607 cm -1 are assigned to the Cu-O stretching frequency in Cu 2 O. 34,35 The nanocrystals were exposed to NaBH 4 to investigate the reduction of the Cu I present in the Cu 2 O to metallic Cu 0 .
Furthermore, since the model catalytic type reactions that will be tested is the reduction of nitrophenols with NaBH 4 , the influence of the NaBH 4 on the Cu-oxide nanocrystals is important. The same experimental conditions as during the catalysis were applied except that the nitrophenols were absent (10 ml of a 6.25 mM NaBH 4 solution was added to a 70 ml suspension of the Cu-oxide nanocrystals (10 mg)). The SEM image indicates that individual cubic structures were maintained. However, the well-defined edges had been smoothed (Figure 4 (A)). As for the chemical composition (as measured by XPS, see Figure 4 (B) for the Cu 2p 3/2 area), the Cu I converted 100% to metallic Cu 0 with the new binding energy at 932.56 eV.
The heterogeneous catalytic activity of the truncated cubic Cu-oxide nanocrystals, with sodium borohydride (NaBH 4 ), was evaluated for the reduction of three different nitrophenols. The nitrophenols under investigation were 2-, 3-and 4-nitrophenol. The time-dependent absorption curves ( Figure 5) show a decrease in the absorption at the peak maxima in the visible region. Manipulation of the data resulted in the construction of the absorption vs time (at the wavelength of peak maxima) and ln(A/A 0 ) vs time graphs ( Figure 5). The later graph exhibited a linear relationship, having a negative slope. The absolute value of the slope is the apparent pseudo-first-order rate constant (see Table 2 for the average of the experiment done in triplicate), 36 which will be used to compare the reduction rates of the different nitrophenols. The order of reactivity from slowest to fastest is:

4-Nitrophenol < 3-Nitrophenol < 2-Nitrophenol
Comparison of the apparent pseudo-first-order rate constant, k', with the Hammett constants (an empirical value relating the reaction rates and equilibrium constants of meta-and parasubstituted aromatic compounds) 37 of the different nitrophenol revealed a linear relationship (see Table 2 for the data and Fig.  5S in the Supplementary Information for the graph). This relationship showed that an increase in the Hammett constant is associated with an increase of k'. A higher Hammett constant is associated with a stronger electron-withdrawing effect (implying more electron density on the nitro-group). This

Figure 5
The UV-Vis absorption spectra over time, the graph of absorption vs time at the given wavelength and the graphs of the ln(A/ A 0 ) vs time (from which the apparent rate constants were calculated for comparison) for the catalytic reduction of 2-, 3-and 4-nitrophenol over the truncated cubic Cu-oxide nanocrystals using NaBH 4 as the reducing agent.
ReseaRch aRticle E. Erasmus 109 S. Afr. J. Chem., 2021, 75, 106-110 https://journals.co.za/content/journal/chem/ effect suggests that the larger electron density causes faster reduction rates of the nitrophenol (over the truncated cubic Cu-oxide nanocrystals using NaBH 4 as the reducing agent). These reduction rates are due to the easier nucleophilic attack of the oxygen (of the nitro group) on the H + , caused by the increased electron density. This attack trend is similar to what was obtained for the electrochemical reduction of nitrophenol with the nitro groups in different positions. 38 The apparent pseudo-first-order rate constants halved when the recovered catalyst (truncated cubic Cu-oxide nanocrystals) was used for a second catalytic cycle. After one catalytic cycle, the recovered catalyst's SEM image revealed that the particles retain their global cubic structure, but the planes and edge are no longer smooth (Figure 6, (A)). This reduced smoothness corresponds to the surface degradation of the truncated cubic Cu-oxide nanocrystals exposed to NaBH 4 (see Figure 4). This loss in the catalytic activity of the truncated cubic Cu-oxide nanocrystals correlates to the literature 39 , where it was reported that cubic Cu-oxide nanocrystals showed a significant loss in catalytic activity (for an oxidative arylation reaction) after the first cycle, from 94% to 47% yield. Li et al. 39 also found the degradation of the straight edges attributed to the "dissolution and reconstruction of active atoms at corners or edges during catalysis. In analogue to this, it is proposed that during the catalytic reduction of nitrophenol, some of the active Cu-oxide fragments break away from the truncated cubic Cu-oxide nanocrystals. Subsequently, these smaller particles aggregate randomly to the large Cu-oxide particle. The decreased activity of truncated cubic Cu-oxide nanocrystals "could be due to the loss of active atoms into the reaction mixture." After the second catalytic cycle, the SEM image showed that needle-like nanostructures protrude from the surfaces of the rough "cubic" structures ( Figure 6 (B)). XPS measurements of the recovered catalysts revealed that only metallic Cu 0 (binding energy Cu 2p 3/2 = 932.56 eV) was present in the recovered catalyst sample. Similar to when the truncated cubic Cu-oxide nanocrystals were treated with NaBH 4, and there was no evidence of Cu I of Cu II species in the materials.

Conclusion
Herewith we have disclosed the successful preparation of truncated cubic Cu-oxide nanocrystals with a ca. 600 nm edge length. The detailed morphologies of the polyhedral structures were characterised by XRD and SEM, while the chemical and physical properties were analysed by UV-Vis, ATR FTIR, and XPS.
The XRD pattern indicates that the truncated cubic Cu-oxide nanocrystals are not exclusively present in the Cu I state but that some Cu 0 is also present. The XPS confirmed this presence, revealing some Cu II present in the Cu(OH) 2 form. The structure and chemical composition are unstable towards exposure to the reducing agent NaBH 4 . Treatment with NaBH 4 reduced the Cu I fully to metallic Cu 0 . The sharp and well-defined facets of the truncated cubic structure's edge and corners were lost and became round.
The truncated cubic Cu-oxide nanocrystals can be used as a catalyst for reducing a variety of different nitrophenols. The order of reactivity from slowest to fastest is: 4-Nitrophenol < 3-Nitrophenol < 2-Nitrophenol A directly proportional relationship was obtained between the pseudo-first-order rate constant, k', and the Hammett constants of the different nitrophenol. This relationship implies that an increase in the electron-withdrawing ability of the nitrophenol (higher Hammett constant) results in a faster reaction (an increase of k').
The recovered catalyst was found to be in the metallic Cu 0 state, which was not as active as the Cu 2 O state. The structure deformed from the truncated cube to a "fussy" cubic structure after one cycle. After a second cycle, the SEM revealed that needle-like nanostructures protrude from the surfaces of rough "cubic" structures.