Ignacio
Chamorro-Mena
,
Noemi
Linares
* and
Javier
García-Martínez
*
Laboratorio de Nanotecnología Molecular, Departamento de Química Inorgánica, Universidad de Alicante, Ctra. San Vicente-Alicante s/n, E-03690 San Vicente del Raspeig, Spain. E-mail: noemi.linares@ua.es; j.garcia@ua.es; Web: https://www.nanomol.es
First published on 4th September 2023
This study describes how the optimization of Cu2O/CuO heterostructures can enhance their (photo)catalytic performance. More specifically, the evaluation of catalysts with different Cu2O/CuO molar ratios was used to optimize their performance for the hydrogenation of 4-nitrophenol under both blue-LED light and dark conditions. For the first time, we analyzed the effect of blue LED irradiation on this reaction and found that when blue LEDs are used as the light source, a Cu2O/CuO ratio of 0.15 results in rate constants 7 to 3 times higher than those of catalysts with either lower (0.01) or higher (0.42) ratios. Furthermore, this photocatalyst shows good stability, >70% after 5 cycles, and excellent chemoselectivity in the selective reduction of the nitro group in the presence of other functionalities, i.e. –COOH, –CONH2 and –OH.
Efforts to develop catalytic processes that are more sustainable and cost-effective are essential from both economic and environmental standpoints. In this sense, utilizing photocatalytic synthetic routes, which rely on light as an energy source to drive chemical reactions, can offer a more sustainable alternative. Since the 1970s, photocatalysis has been extensively researched and applied in various fields, including water splitting, air and water purification, N2 fixation, and CO2 reduction, among others.12 Titanium dioxide (titania) is one of the most versatile and effective semiconductor photocatalysts used in a variety of reactions. It has shown almost quantitative yields of the corresponding anilines for basic nitroarenes, making it an excellent option for this reaction.13,14 However, using titania as a photocatalyst has a significant drawback – its low activity in the visible spectral range. Therefore, one needs to use ultraviolet light sources or dope the titania surface with metals and/or organics, which can substantially increase the procedure's cost and complexity. Recently, the use of other semiconductors that are less photocatalytically active than titania but respond in the visible range has been explored.15–17 Examples of such semiconductors are CuO and Cu2O, which have been tested as photocatalysts in various reactions.15,16,18 However, despite their advantages such as low cost, versatility, non-toxicity and good response under visible light, they have not yet been evaluated for the photoinduced hydrogenation of nitroarenes to the best of our knowledge.
Here, we explore the use of Cu2O/CuO heterojunctions, obtained via the controlled calcination of the Cu(BTC) MOF (or HKUST-1), for the photocatalyzed hydrogenation of nitroarenes. More specifically, we evaluated a series of Cu2O/CuO materials with different molar ratios at 25 °C under both dark and blue-LED light (6 W) conditions using NaBH4 as a hydrogen donor to assess the role of visible light in the enhancement of the rate constant of the reaction (see Fig. 1 for a schematic representation). Under both dark and irradiation conditions, we observed a maximum in the performance of heterojunctions with intermediate Cu2O/CuO ratios. The use of light is able to enhance the catalytic activity even further, provoking a 2-fold increase in the rate constant compared to the experiments under dark conditions.
Fig. 1 Schematic representation of the photocatalytic reduction of nitroarenes into aminoarene derivatives by using Cu2O/CuO catalysts derived from the calcination of the Cu(BTC) MOF. |
The crystallinity of the materials was characterized by X-ray powder diffraction (XRD) in a SEIFERT 2002 apparatus from 20 to 80° 2θ using a scanning velocity of 1° min−1 and CuKα (1.5418 Å) radiation. We analyzed the diffractograms using the X'Pert HighScore Plus software package, which enabled us to identify and quantify the various phases obtained by comparing them with the standard Powder Diffraction Files of the International Center for Diffraction Data (ICCD).
The evolution of the Cu2O/CuO ratio on the catalysts’ surface was studied by X-ray photoelectron spectroscopy (XPS). These measurements were carried out using VG-Microtech Multilab instrument, with MgK-alpha radiation at an energy of 1253.6 eV and a pass energy of 50 eV. The analysis pressure during data acquisition was 5 × 10−7 Pa. The deconvolution of the spectra was made and the obtained areas under the peaks were estimated by calculating the integral of each peak after subtracting a Shirley background and fitting the experimental peak to a combination of Lorentzian/Gaussian lines with a proportion of 30% to 70%. Binding energies were referenced to the C 1s line at 284.6 eV from adventitious carbon.20
Diffuse reflectance ultraviolet-visible (DRUV) spectroscopy was used to study the optical properties of the copper oxides obtained from the MOF calcination. DRUV spectra were obtained in air, at room temperature, and in the range 1200–200 nm using a Shimadzu UV-2401 PC spectrophotometer. BaSO4 was used as the reference material.
The UV-Raman spectra were obtained to further monitor the changes in the crystalline structure. The analyses were carried out on a Jasco NRS-5100 dispersive Raman spectrometer using a 632 nm He–Ne laser source.
The porosity of the catalysts was evaluated by N2 physisorption at 77 K. The adsorption/desorption isotherms were obtained in a Quadrasorb-Kr/MP apparatus from Quantachrome Instruments. The samples were previously degassed for 12 h at 120 °C at 5 × 10−5 bar. Adsorption data were analyzed using QuadraWinTM (version 6.0) software from Quantachrome Instruments. BET surface areas were calculated using the BET equation.
The chemoselectivity of the best perfoming catalytic system was evaluated by reducing nitroarenes containing diverse functional groups. An identical procedure was used to monitor the hydrogenation of the various nitroarenes, and a mass spectrometer was used for the analysis. Specifically, an Agilent Model 1100 Series high-performance liquid chromatograph coupled simultaneously with a variable wavelength visible-UV detector and an Agilent Model 1100 Series LC/MSD Trap SL mass spectrometer with an ion trap analyzer and MS/MS capability were used, with a mass range of 50 to 3000 uma.
Fig. 3 FESEM micrographs of (A) Cu(BTC), (B) Cu2O/CuO = 0.42, (C) Cu2O/CuO = 0.15 and (D) Cu2O/CuO = 0.01. |
The XPS spectra shown in Fig. 2B confirm a similar Cu2O/CuO evolution on the surface of the catalysts. The Cu 2p3/2 spectra reveal a decrease in both the area and size of the peak corresponding to Cu(I) in the lattice (932 eV) as the temperature increases. Meanwhile, the peak corresponding to lattice Cu(II) (933 eV) becomes larger. In all cases, there is also a small peak at 936 eV due to Cu bound to –OH groups, which is usually attributed to water adsorbed on the surface of this material.24 The diffuse reflectance UV-vis (DRUV) spectra of the solids indicate similar optical absorption properties in all the Cu2O/CuO photocatalysts produced by thermal oxidation of Cu(BTC) (Fig. 2D). In all cases, a broad absorbance region spanning in the 250–1200 nm range was observed, with an absorption edge at approximately 850 nm for all the Cu2O/CuO heterostructures. These results suggest that the materials possess excellent light harvesting properties in the visible light range.
As shown in Fig. 3, the FESEM micrographs clearly demonstrate that there are significant changes in the morphology and texture of the samples as the calcination temperature increases. Initially, the materials consisted of perfectly octahedral Cu(BTC) particles with a smooth surface (Fig. 3A). After calcination, the particles retained their octahedral morphology, but their texture became much rougher and more porous due to the formation of oxide particles, as seen in the FESEM micrographs (Fig. 3B, sample calcined at 300 °C, Fig. 3C, sample calcined at 400 °C and Fig. 3D, sample calcined at 350 and 450 °C). In order to further analyze these changes, N2 physisorption at 77 K was employed to study the textural properties of the catalysts, while TEM was used to analyze both the size and morphology of the primary nanoparticles that form the larger octahedra. As observed in Fig. S1,† the size of the primary nanoparticles is also highly affected by the temperature of treatment. The material calcined at 300 °C (Cu2O/CuO = 0.42) shows the smallest particle size of around 9–15 nm, and the particle size increases with the temperature of calcination and, subsequently, the CuO content. Consequently, the Cu2O/CuO = 0.15 catalyst presents a particle size between 25 and 50 nm, while the Cu2O/CuO = 0.01 catalyst has the largest particle size, with a particle size between 45 and 100 nm. Similar to what was also observed in other studies,25 the particle size follows an opposite trend of the Cu2O content, the higher the Cu2O content, the smaller the particle size. However, these differences in particle size are not reflected in the textural properties. As shown in the N2 physisorption isotherms at 77 K (Fig. S2†), the adsorption of the MOF material drastically drops after its calcination. All the catalysts display very low adsorption profiles, indicating a very low porosity with BET surface areas of around 10 m2 g−1 in all cases.
The catalytic activity of the materials was assessed for the reduction of 4-nitrophenol to 4-aminophenol using NaBH4 at 25 °C, both with and without irradiation by blue LEDs (see the conversion of the different materials in Fig. S3–S5†). The rate constants were obtained from the slopes of the ln(C0/C) versus time plot, which corresponds to pseudo-first-order kinetics. Fig. 4A represents the experiments conducted under blue-LED irradiation, while Fig. 4B compares those results with the non-irradiated experiments (see Fig. S3–S5† for all the ln(C0/C) versus time plots).26
The sample obtained by calcination at 400 °C, which has an intermediate Cu2O/CuO molar ratio of 0.15, exhibits the best performance under both light and dark conditions, suggesting an optimal catalyst composition for this reaction. Interestingly, the use of the sample obtained by calcination at 400 °C exhibits a three-fold increase in the rate constant under dark conditions compared to the sample with either a higher Cu2O/CuO ratio (calcined at 300 °C, 0.42) or a lower Cu2O/CuO ratio (calcined at 350/450 °C with a Cu2O/CuO ratio of only 0.01). To understand the origin of this optimal composition, XPS analyses were performed on the catalysts after one reaction cycle (Fig. S6†). During catalysis, the surface of the materials is modified with a highly reducing agent such as NaBH4, which leads to the formation of Cu2O and/or Cu. These two oxidation states are difficult to distinguish by the Cu 2p3/2 XPS spectra. Furthermore, a peak at 933.8 eV, corresponding to CuO, is also detected in all three samples, albeit in much lower amounts compared to the initial catalysts. As expected, the addition of NaBH4 to the reaction mixtures caused a significant change in the Cu2O/CuO ratios on the surface of the catalysts. This was due to the reduction of both oxides, which led to the formation of Cu(0) through the reduction of Cu2O, while at the same time, Cu2O was regenerated as a result of the reduction of CuO. The presence of Cu(0) and the reduction of oxygen in the lattice were confirmed by XPS (see the Cu LMM spectra in Fig. S7† and the O 1s spectra in Fig. S8† and their descriptions below). These results indicate that a small fraction of Cu(0) is generated in situ on the surface during catalysis. The in situ formation of these highly active Cu(0) species during the reaction has been reported elsewhere.27 In those studies, it was concluded that these are the active species for the nitroarene reduction. On the other hand, the presence of CuO after reaction in this reductive medium is unexpected; however, the oxygen dissolved in the aqueous medium and/or the high alkalinity of the solution may contribute to the oxidation of Cu(0) or Cu2O to CuO, as described elsewhere.28 Under the reaction conditions, catalysts present a combination of the three states Cu(0), Cu(I) and Cu(II) in different proportions, which are related to their initial composition. The best performing catalyst should have a balanced Cu/Cu2O/CuO ratio that favors the formation and cyclic regeneration of the active species, which is Cu(0), in order to achieve an optimal rate constant value. Furthermore, the literature extensively describes how the increased number of oxygen vacancies (Ov) in these catalysts can enhance nitroarene adsorption, ultimately improving the catalyst's performance.10 To quantify the number of oxygen vacancies (OV) present in our catalysts, we analyzed the O 1s spectra of all materials before and after the reaction. The broad O 1s bands were deconvoluted into three peaks (Fig. S8†). The peak at ca. 529.7–530.3 eV can be assigned to the lattice oxygen (OL) in the Cu2O/CuO materials. The second peak at 531.2 eV has been described as the result of the ionization of oxygen species due to the compensation of the oxygen deficiencies in the copper oxides, so it can be related to the oxygen vacancies (OV).24 The last peak, centred at ca. 533.0 eV, is caused by weakly adsorbed oxygen species (OA), attributed to water adsorbed on the catalyst. The relative concentrations of the various oxygen species on the materials’ surface were calculated by determining the area under the corresponding curve (Table S1†). Interestingly, the percentage of OV in the best performing catalyst is slightly higher than that in the other solids, even after the reconstruction of the catalysts’ surface that occurs during the reaction, which may contribute to its superior activity. It is worth noting that even after the reaction, the OL signal decreases but does not completely disappear. This observation confirms the presence of Cu2O/CuO in the highly reductive reaction medium.
When comparing rate constants using the same catalysts under light and dark conditions, it is evident that the presence of Cu2O plays a critical role in enhancing the catalytic performance (as depicted in Fig. 4B). Interestingly, regardless of the initial Cu2O/CuO ratio in the catalyst (0.15 or 0.42), the rate constant of reactions carried out under blue-LED light exhibits a two-fold improvement compared to the same experiment performed under dark conditions. The significant variation in activity observed between blue-LED irradiation and dark conditions is apparent from the gradual reduction in the maximum absorption of the nitrophenolate ion (as illustrated in Fig. 4E and F). Using Cu2O/CuO heterojunctions, the disappearance of the UV-band occurs within 9 minutes in the presence of light, whereas the process slows down considerably to 25 minutes in the dark. In contrast, the sample mainly composed of CuO (Cu2O/CuO = 0.01) exhibits similar rate constants with and without irradiation, with values significantly lower than those obtained from nitro photoreduction with the other two samples. Here, it is worth highlighting that Cu2O/CuO materials have previously shown improved catalytic performances compared to those shown for a single phase of CuO or Cu2O for other reactions, i.e. C–N coupling,29 CO oxidation,30,31 trichlorosilane synthesis,32 and NO reduction,33 among others. Although the use of Cu2O/CuO heterojunctions has been previously described for the catalytic hydrogenation of nitroarenes,34–36 to the best of our knowledge, this study is the first example of the use of these Cu2O/CuO heterojunctions for the photocatalytic reduction of nitroarenes.
Various mechanisms have been proposed for the reduction of nitrophenol to aminophenol using borohydride as a reducing agent.37 Most of these mechanisms describe the initial adsorption of the reducing agent on the catalyst's surface and the subsequent transfer of hydrogen ions/atoms or electrons from the borohydride to the nitrophenolate ion, induced by the catalyst. This mechanism is similar to what occurs in the hydrolysis of borohydride. So, analyzing the effect of blue-LED light irradiation on the borohydride hydrolysis could provide valuable insights into the influence of this factor on the photocatalyzed reduction of nitroarenes under the studied conditions.38 Therefore, we evaluated the best performing catalyst, i.e., the Cu2O/CuO = 0.15 catalyst, in sodium borohydride decomposition, under blue-LED light and dark conditions, by measuring the evolution of hydrogen. Fig. S9† shows that experiments conducted under light decompose sodium borohydride two times faster than experiments conducted in the dark, which can explain the beneficial effect of light in the nitrophenol reduction as well. Similar to the observations made during borohydride hydrolysis, the rate constants for the reduction of 4-nitrophenol are also twice as large in the presence of light as compared to the processes conducted without light. The positive effect of light on the hydrolysis of borohydrides has also been observed for other photocatalysts, leading to similar conclusions.39–41
Furthermore, the stability and reusability of the best performing catalyst (Cu2O/CuO = 0.15) were evaluated under blue-LED light conditions, as shown in Fig. 4D. The material consistently demonstrated excellent performance for five consecutive cycles without a significant loss in activity, maintaining conversion rates of at least 70% throughout. In order to determine the stability of the catalyst over cycles, we have analyzed the catalyst's surface composition after each photocatalytic cycle (Fig. S10†). As shown by the XPS Cu2p3 spectra, the percentages of Cu(I) and Cu(II) do not change drastically so, we can conclude that this catalyst is not only reusable, but also quite stable for at least 5 cycles. However, it is worth noting that the cycles in which the conversion decreases (cycles 2 and 3) are associated with an increase in the Cu(II) content; however, when the Cu(II) content decreases again, the conversion increases. As previously described, under our reaction conditions, the catalysts present a combination of the three states Cu(0), Cu(I) and Cu(II) in different proportions. Thus, the conversion rate depends on the Cu/Cu2O/CuO ratio, with higher amounts of CuO leading to lower rates. This is the same conclusion that can be drawn from the recycling experiments.
We also performed four control experiments, specifically, (1) a blank experiment, without a catalyst, which did not yield any appreciable amount of 4-aminophenol (Fig. 4C). This observation confirms that NaBH4 alone cannot reduce 4-nitrophenol under the studied conditions, which is likely due to the high kinetic barrier of this reaction.42 (2) The evaluation of a commercial CuO sample as a benchmark catalyst. This experiment resulted in lower performance compared to our catalysts (Fig. 4C), indicating the high catalytic activity of the MOF-derived solids. (3) An experiment without the use of NaBH4, which showed no conversion of 4-nitrophenol. This experiment shows the need for the presence of an agent capable of contributing hydrogen to the catalytic system in order to achieve the reduction of the nitro group. In our case, this agent is sodium borohydride, but other compounds can be used, such as formic acid, which as can be seen would also be crucial for the reduction to take place.43,44 (4) With the aim of obtaining a higher proportion of Cu2O in the Cu2O/CuO catalyst, we also evaluated the calcination of the MOF at a lower temperature (250 °C), which was insufficient to cause the oxidation of the MOF.
Finally, to assess the selectivity of the best performing catalyst (Cu2O/CuO = 0.15) towards the reduction of nitro groups in the presence of other functional groups, various substituted nitroarenes were tested. As shown in Table 1, selectivities nearly 100% were obtained in every case at very high conversions (>70%). Entries 1 and 2 demonstrate that the presence of a hydroxy group results in similar conversion values for both isomers, regardless of the relative position of the ring substituents (ortho or para). The excellent performance of our photocatalyst is evidenced with these examples, where almost quantitative conversions are obtained in a short time (9 min) and under mild conditions, even taking into account that the substitution of an electron donor group (as the –OH) at the ortho and para positions can increase the electron density at the nitro group by the resonance effect, which should hinder the reduction of the nitro group.45 In the case of substituents with carbonyl functional groups, the easily reducible functionalities, such as –COOH and –CONH2 (entries 3 and 4), remained unaffected.46 However, as shown in Table 1, the conversion falls moderately with respect to that of nitrophenols.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d3dt01670f |
This journal is © The Royal Society of Chemistry 2023 |