Dual-Function Bare Copper Oxide (Photo)Catalysts for Selective Phenol Production via Benzene Hydroxylation and Low-Temperature Hydrogen Generation from Formic Acid

Abstract

In this work, bare copper oxide-based catalysts were synthesized and evaluated for their dual (photo)catalytic activity in two model reactions: hydrogen generation via formic acid decomposition (FAD) and the photocatalytic hydroxylation of benzene to phenol. Catalysts were prepared from copper nitrate and copper acetate precursors and calcined for either 10 min or 2 h. Their structural and surface properties were characterized by wide-angle X-ray diffraction (WAXD), Raman spectroscopy, and BET surface area analysis. FAD was conducted under mild thermal conditions and monitored via ¹H NMR spectroscopy. At the same time, benzene hydroxylation was performed under UV irradiation and analyzed by gas chromatography (GC) and high-performance liquid chromatography (HPLC). All synthesized catalysts outperformed commercial CuO in the selective oxidation of benzene. The nitrate-derived sample calcined for 10 min (NCuO 10 min) achieved the best performance, with a phenol yield of ~10% and a selectivity of up to 19%, attributed to improved surface properties and the presence of Cu(I) domains, as indicated by Raman spectroscopy. For FAD, complete conversion of formic acid was achieved at low temperatures, with selective H₂ and CO₂ evolution and complete suppression of CO, even under short reaction times and low catalyst loadings. These results demonstrate the potential of nitrate-derived CuO catalysts as versatile, dual-function materials for sustainable applications in selective aromatic oxidation and low-temperature hydrogen generation, without the need for noble metals or harsh conditions.

1. Introduction

The development of catalytic systems capable of operating under mild conditions has become a central goal in the pursuit of sustainable and energy-efficient chemical processes. Traditional catalytic approaches for key industrial reactions, such as hydrogen generation and aromatic hydroxylation, typically require high temperatures, elevated pressures, or harsh reagents, resulting in significant energy consumption and environmental impacts [1,2,3,4,5]. Therefore, the design of catalysts that are both efficient and functional under mild, eco-friendly conditions is essential for advancing greener chemical technologies.

In this context, copper oxide-based photocatalysts have gained considerable attention. Both Cu₂O and CuO are p-type semiconductors. When combined to form a CuO–Cu₂O heterojunction, the interface promotes effective charge separation and suppresses electron–hole recombination, thereby enhancing photocatalytic activity [6,7,8,9]. Moreover, Cu-based semiconductors such as Cu₂O and CuO are earth-abundant, low-cost, and environmentally benign, in contrast to precious-metal catalysts, whose mining and processing generate high carbon footprints and costs [10,11]. These properties make CuO–Cu₂O heterostructures particularly appealing for scalable and sustainable catalytic applications [12,13].

Despite their considerable potential, bare copper oxides have been relatively underexplored for catalytic applications under ambient conditions—particularly in the context of complex organic transformations and to promote hydrogen production [14,15,16,17,18,19,20]. In this study, we investigate the dual-function (photo)catalytic behavior of a series of copper oxide materials through two representative reactions carried out under mild conditions—(i) the selective photocatalytic hydroxylation of benzene to phenol under UV light irradiation at near-room temperature and (ii) hydrogen generation via formic acid decomposition (FAD), conducted with low catalyst loading and short reaction times. The combined results underscore the practical relevance of these materials as efficient, low-cost, and environmentally sustainable platforms for dual-purpose catalytic applications.

The photocatalytic hydroxylation of benzene to phenol using hydrogen peroxide (H₂O₂) as the oxidant has emerged as a sustainable and selective alternative to the traditional cumene process, which typically requires harsh conditions and generates significant by-products [21,22]. H₂O₂ is widely regarded as a green oxidant due to its high oxidative potential, its decomposition into benign products (H₂O and O₂), and its compatibility with mild reaction conditions [23]. Various semiconductor-based photocatalytic systems have demonstrated promising performance in this transformation. For example, He et al. reported that surface-modified Cu₂O nanoparticles supported on defect-rich graphene achieved a phenol selectivity of 64% at ~30% benzene conversion, attributed to the hydrophobic interaction favoring benzene adsorption over phenol [24]. Mancuso et al. have also reported the key role of hydrophobic supports for improving selectivity in photocatalytic benzene hydroxylation [25,26,27]. An efficient system was also demonstrated using Ni–CuWO₄/g-C₃N₄ nanocomposites: under solar light and in the presence of H₂O₂, benzene conversion reached 98.5%, with phenol selectivity of 82.7% and phenol yield of 81.5% in just 15 min. This performance was attributed to favorable band alignment and enhanced visible-light absorption, which promoted efficient charge separation and minimized recombination [28]. More recently, Cu/N–TiO₂ composites (Cu oxide supported on N-doped TiO₂) were synthesized and tested under visible light irradiation with H₂O₂. Cu/N–TiO₂ delivered a phenol yield of approximately 25% after 360 min of irradiation, significantly outperforming Fe/N–TiO₂ and V/N–TiO₂ under the same conditions. The enhanced performance was attributed to improved visible-light absorption due to Cu–TiO₂ interactions and low phenol affinity at the catalyst surface, minimizing phenol overoxidation reactions [29]. Overall, although copper oxide-based composites and heterojunction systems have been widely investigated, studies focusing on unmodified or bare copper oxides for the selective photocatalytic hydroxylation of benzene remain relatively scarce and insufficiently explored.

Formic acid (FA) is considered an ideal hydrogen storage material due to its high hydrogen content and its stability in liquid form under ambient conditions [30,31,32,33,34,35]. Hydrogen can be efficiently released from FA via catalytic dehydrogenation. (HCOOH → H₂ + CO₂, ΔG⁰ = −32.9 kJ·mol⁻¹). However, a competing dehydration pathway can also occur, resulting in the formation of carbon monoxide (CO), a highly undesirable by-product due to its ability to poison metal catalysts and hinder the practical application of FA in fuel cell technologies (HCOOH → H₂O + CO, ΔG⁰ = −12.4 kJ·mol⁻¹ [36]. When appropriate catalysts are employed, selective FA dehydrogenation can be achieved with minimal CO generation. Among the most studied systems, heterogeneous noble metal catalysts have demonstrated high thermal stability and low CO formation during FA dehydrogenation (FAD) [37,38,39,40,41,42,43,44]. Nevertheless, the scarcity and high cost of noble metals have prompted significant efforts toward the development of cost-effective alternatives based on earth-abundant transition metals. Catalysts derived from Fe, Ni, and Co have been explored, but they generally exhibit low catalytic activity and poor selectivity in FAD [45,46,47].

Copper, being significantly less expensive than noble metals, represents an attractive alternative. However, only a limited number of Cu-based catalysts have been reported for this transformation [48,49,50]. The few existing examples achieve appreciable FA conversion only at elevated temperatures (>200 °C), primarily due to challenges in stabilizing the active Cu species within the support matrix and the tendency of Cu nanoparticles to aggregate under reaction conditions. In light of these limitations, we investigated the performance of our bare copper oxides as potential catalysts for FAD.

The catalysts investigated in this work were synthesized from copper nitrate and copper acetate precursors and extensively characterized using wide-angle X-ray diffraction (WAXD), Raman spectroscopy, and Brunauer–Emmett–Teller (BET) surface area analysis. Catalysts synthesized from copper acetate are referred to as ACuO, while those derived from copper nitrate are denoted as NCuO. Each was prepared using two different calcination times (10 min and 2 h). Their catalytic performance was assessed under ambient conditions in two distinct reactions: photocatalytic hydroxylation of benzene to phenol under UV irradiation and formic acid decomposition in the dark.

2. Results and Discussion

2.1. Catalysts Characterization

Wide-angle X-ray diffraction (WAXD) analysis was carried out on all catalyst samples to investigate how different calcination times affect their crystalline structure (Figure 1).

WAXD analysis confirmed the formation of monoclinic CuO (tenorite) as the dominant and possibly exclusive crystalline phase in all catalyst samples. The most intense diffraction peaks observed at 2θ ≈ 35.5° and 38.7° correspond to the (–111) and (111) planes of CuO, consistent with standard reference patterns (JCPDS 45-0937) [51]. Additional reflections observed at lower (e.g., 32.5°) and higher angles (e.g., 48.6°, 53.5°, 58.3°, 66.2°, and 72.4°) further support the presence of highly crystalline CuO [52]. In contrast, no clear evidence of Cu₂O (cuprite) was detected in the WAXD patterns. The characteristic reflections of Cu₂O at 2θ ≈ 36.4° (111) and 61.3° (220) [53] were either completely absent or indistinguishable from the background noise. This suggests that, if present, Cu₂O is below the detection limit of the WAXD technique—either due to low crystallinity, small particle size, or trace amounts.

Table 1. Summary of structural parameters of all catalyst samples, including precursor type, calcination time, estimated crystallite sizes based on WAXD peak broadening at 2θ = 35.5°, and specific surface area (SSA) obtained by BET analysis.

Samples	CuXO Precursor	Calcination Time	Crystallite Size, nm	SSA, m2/g
NCuO 10 min	Copper nitrate	10 min	16.9	6
NCuO 2 h	Copper nitrate	2 h	18.1	6
ACuO 10 min	Copper acetate	10 min	18.6	5
ACuO 2 h	Copper acetate	2 h	19.9	4
Commercial CuO	-	-	16.7	2

summarizes the structural parameters of the synthesized CuO–Cu₂O catalysts, including CuO crystallite sizes estimated via the Scherrer equation 2θ = 35.5°, along with the specific surface area (SSA).

Table 1 summarizes the crystallite sizes (estimated from WAXD peak broadening at 2θ = 35.5°) and specific surface areas (SSAs) of the synthesized catalysts and commercial CuO. The data show that NCuO samples, derived from copper nitrate, tend to exhibit slightly smaller crystallite sizes (16.9–18.1 nm) compared to their ACuO counterparts, which reach up to ~19.9 nm after 2 h of calcination. Interestingly, the specific surface area remains at 6 m²/g for both NCuO 10 min and 2 h, while it progressively decreases from 5 to 4 m²/g in ACuO samples as the calcination time increases. This trend is consistent with the growth of larger, more compact CuO crystallites under prolonged thermal treatment.

However, the slightly smaller crystallite sizes and higher surface areas in the NCuO series suggest a greater degree of structural disorder or nanoscale dispersion, potentially linked to the formation of highly defective CuO or the partial presence of Cu(I) species. These structural features, particularly the reduced crystallite size and higher SSA observed in NCuO samples, are consistent with the presence of increased lattice disorder and potentially defect-related phenomena. As will be further discussed in the Raman analysis section below reported, these structural variations may be correlated with distinct vibrational features and spectral shifts, providing additional evidence for differences in phase purity and local bonding environments among the various catalyst systems.

Raman spectroscopy was performed to explore the structural and vibrational distinctions of CuO catalysts derived from nitrate (NCuO) and acetate (ACuO) precursors (calcined for 10 min and 2 h), compared with commercial CuO (Figure 2).

All samples display the CuO-characteristic phonon modes: strong Ag (~280–290 cm⁻¹) and Bg (~330–350 cm⁻¹) bands, plus a broader Cu–O stretching feature around ~620 cm⁻¹, consistent with pure CuO reported in earlier works [54,55]. Significantly, the NCuO 2 h sample exhibits a distinct ~218 cm⁻¹ band, absent in the ACuO series and commercial CuO. This feature corresponds to the T_2g(2) mode of Cu₂O, a well-known Raman marker for Cu(I) species even in poorly crystalline or surface-localized phases [56,57]. A weaker analog of this band also appears in NCuO 10 min, implying early-stage Cu₂O formation that diminishes yet persists partially after extended calcination. In contrast, ACuO samples show no detectable Cu₂O-related band, indicating that CuO remains largely stoichiometric and free of Cu(I). Further evidence emerges from the behavior of the CuO Ag mode below 300 cm⁻¹. In NCuO samples, particularly NCuO 2 h, this band is noticeably red-shifted and broadened, reflecting lattice strain, oxygen vacancies, or interfacial distortions potentially arising from CuO–Cu₂O interactions or partial reduction [58,59]. Even ACuO 10 min exhibits a mild red-shift, albeit without Cu₂O band manifestation, suggesting minor surface disorder or nascent Cu(I) presence insufficient to generate a resolvable Raman band. Moreover, in NCuO samples, the CuO-associated band in the 600–650 cm⁻¹ region exhibits a red-shift (centered around 630–635 cm⁻¹) and increased width compared to both ACuO and commercial CuO. Such behavior is consistent with structural disorder, oxygen vacancies, or the presence of Cu₂O-related vibrational modes activated in defective or interfacial environments. Similar observations have been reported in mixed-phase CuO/Cu₂O systems, where the 630 cm⁻¹ band may include contributions from multi-phonon overtones or T_1u modes of Cu₂O that become Raman-active due to local symmetry breaking or defectivity [60,61].

To further investigate potential structural differences between the catalysts, FTIR spectra were recorded for all samples (Figure S1a,b, Supplementary Materials). In the 450–700 cm⁻¹ region, typically associated with Cu–O lattice vibrations [62], among the nitrate-derived samples (Figure S1a), the NCuO 2 h spectrum shows a broader and more intense absorption band compared to NCuO 10 min and commercial CuO. This band broadening and red shift suggest a greater degree of structural disorder, likely due to the presence of defective CuO environments, oxygen vacancies, or Cu(I)–Cu(II) interfacial interactions. These spectral features are consistent with the Raman results, where NCuO 2 h showed a red-shift and broadening of the main CuO band (~290 cm⁻¹), which is typically activated or distorted in the presence of lattice strain and interfacial defects involving Cu₂O domains. For the acetate-derived samples (Figure S1b), ACuO 10 min shows the most intense absorption in the Cu–O region, but the band shape is more defined and centered similarly to commercial CuO, suggesting a higher degree of crystallinity and fewer structural defects. The ACuO 2 h sample displays slightly broader bands but still lacks the marked red-shift and broadening observed in the nitrate-derived counterparts, in line with Raman data that showed minimal deviation from bulk CuO vibrational features.

The intrinsic thermal decomposition behavior of the copper precursors plays a pivotal role in the structural differences observed between NCuO and ACuO. Copper nitrate decomposes through multiple stages—starting from hydrated Cu(NO₃)₂·3H₂O to basic copper nitrate intermediates below ∼180 °C, followed by decomposition to CuO, NO₂, H₂O, and O₂ around 200–250 °C [63,64]. During these steps, the release of nitrogen oxides and the catalytic decomposition of NO₂ at CuO surfaces can transiently lower the local oxygen activity, thus generating micro-reducing environments that facilitate partial reduction of Cu(II) to Cu(I), leading to Cu₂O domain formation [65], even when bulk WAXD indicates CuO as the dominant phase. In contrast, copper acetate decomposes more gradually, typically in two main steps: dehydration followed by decarboxylation, mainly yielding CO₂ and H₂O between 190 and 300 °C. This slower and smoother decomposition helps to passivate reactive surfaces and promotes the formation of well-crystallized, stoichiometric CuO nanoparticles with fewer defects and minimal presence of Cu(I) [66]. This mechanistic contrast explains why only NCuO catalysts show Cu₂O Raman features and more pronounced Ag shifts, while ACuO catalysts exhibit clean CuO profiles with fewer structural perturbations.

In summary, Raman and FTIR spectroscopy reveal clear structural differentiation: NCuO materials, especially after 2 h calcination, exhibit evidence of Cu₂O formation and structural disorder, whereas ACuO catalysts retain crystalline, stoichiometric CuO.

2.1.1. Selective Photocatalytic Benzene Hydroxylation to Phenol

The photocatalytic activity of the CuO-based catalysts was evaluated in the selective hydroxylation of benzene to phenol using H₂O₂ as the oxidant under UV irradiation and near-room temperature conditions. All experiments were performed under identical conditions of catalyst loading, light intensity, and substrate concentration in order to compare the intrinsic activity and selectivity of the different materials. Figure 3 and Figure 4, together with

Table 2. Benzene conversion (X_Bz), phenol selectivity (S_phl), CO₂, and other ring-opening compounds’ selectivity (S_Deg), as well as phenol yield (Y_phl) obtained using all the samples under UV light irradiation at the time corresponding to the maximum phenol yield.

Samples	Time to Maximum Phenol Yield	X Bz	Yphl	Sphl	SDeg
NCuO 10 min	60 min	55.5%	10%	18.0%	76.9%
NCuO 2 h	300 min	55%	10.5%	19.0%	75%
ACuO 10 min	180 min	80%	4.5%	5.6%	87.2%
ACuO 2 h	120 min	51.8%	3.7%	7.1%	90.3%
CuO	120 min	89%	2.8%	3.2%	94.5%

, provide a comprehensive overview of the photocatalytic performance of all CuO-based catalysts. Figure 3 shows the evolution of the relative benzene concentration, while Figure 4 presents the corresponding phenol yields over time. These data are summarized in Table 2, which includes the time to obtain the maximum phenol yield, benzene conversion (X_Bz), phenol yield (Y_phl), phenol selectivity (S_phl), and the selectivity toward degradation products (S_Deg), including CO₂ and ring-opening species, recorded at that same irradiation time.

Among all catalysts, NCuO 10 min and NCuO 2 h demonstrate the most favorable compromise between benzene conversion and phenol selectivity. In particular, NCuO 10 min exhibits a faster benzene hydroxylation reaction rate (Figure 3a), reaching its phenol yield maximum (10%) within 60 min. In comparison, NCuO 2 h achieves a slightly higher yield (10.5%) but with a more extended induction period (300 min) (Figure 4a). Notably, these two NCuO samples also show the highest phenol selectivity (18.0% and 19.0%) (Table 2), indicating that they can effectively promote the hydroxylation of benzene while minimizing phenol overoxidation. Although NCuO 10 min exhibits a superior ability to stabilize the generated phenol compared to the other catalysts, a gradual decrease in phenol yield is still observed upon prolonged UV irradiation. This suggests that phenol undergoes further oxidation beyond 60 min, albeit to a lesser extent than in the other systems. This interpretation is supported by the mechanistic analysis derived from HPLC data, which shows the formation of catechol and p-benzoquinone as downstream products (see Scheme 1 and Supplementary Material).

In contrast, the ACuO catalysts (Figure 3b and Figure 4b) and commercial CuO display significantly lower phenol yields (≤7%), despite achieving high benzene conversions (up to 89% for commercial CuO). This is particularly evident in the case of ACuO 10 min, which converts 80% of benzene but yields only 4.5% phenol due to its low phenol selectivity (5.6%) (Table 2).

The comparative photocatalytic benzene hydroxylation performance of the NCuO samples suggests that there is an optimal content of Cu(I)/Cu₂O for achieving a favorable balance between benzene conversion and phenol selectivity. Although NCuO 2 h shows a more intense Raman band at ~218 cm⁻¹ (Figure 2a), indicative of a greater proportion of Cu₂O, it also exhibits slower benzene degradation kinetics (Figure 3) compared to NCuO 10 min, and only marginally higher phenol selectivity. In contrast, NCuO 10 min, which exhibits a weaker but still detectable Cu(I)-related signal, reaches its maximum phenol yield more rapidly and maintains a high selectivity. This suggests that a moderate amount of Cu₂O—possibly forming dispersed CuO–Cu₂O interfacial domains—is sufficient to enhance charge separation and suppress overoxidation, without excessively compromising reaction rates. When Cu₂O becomes more prominent, as in NCuO 2 h, the photocatalytic system likely suffers from lower charge mobility and reduced surface reactivity, as also supported by its larger crystallite size and reduced specific surface area.

This interpretation aligns with literature reports showing that CuO–Cu₂O heterostructures exhibit improved photocatalytic properties due to efficient electron–hole separation across the p–p junction, but that excessive Cu₂O content can hinder charge transport and reduce the number of catalytically active sites. For example, Haldar et al. demonstrated that optimized CuO–Cu₂O interfaces in thin films significantly enhanced charge collection and photocatalytic activity [9]. Similarly, recent studies on composite Cu₂O/CuO systems show that an imbalance in phase composition can lead to poor surface redox reactions [67].

In addition to phenol formation, the reaction pathway includes further oxidation processes leading to ring-opening products and complete mineralization of benzene to CO₂, which are collectively quantified in the degradation selectivity (S_Deg, Table 2). This parameter provides insight into the extent to which the catalyst promotes deep oxidation beyond the desired mono-hydroxylation step. The commercial CuO catalyst exhibits the highest S_Deg value (94.5%), indicating a strong propensity for complete benzene degradation rather than selective transformation to phenol. This behavior is consistent with its large crystallite size and low surface area (as shown by WAXD analysis and BET), which are typically associated with fewer defect sites and a predominance of oxidative pathways. Similarly, the ACuO samples show elevated degradation selectivity (87.2% and 90.3% for ACuO 10 min and 2 h, respectively), which corresponds to their poor phenol selectivity and yield (S_phl < 8%). In contrast, the NCuO materials—particularly NCuO 10 min and NCuO 2 h, display significantly lower S_Deg values (76.9% and 75.0%), indicating a more controlled oxidation process. This is consistent with the Raman evidence of subtle structural disorder and possible Cu(I)/Cu(II) interfaces, which may play a role in modulating radical reactivity and reaction selectivity [8,68,69]. Notably, these catalysts exhibit the best phenol selectivity (18.0% and 19.0%) and yield (~10%), suggesting that a reduced tendency toward ring-opening degradation is essential for achieving selective hydroxylation. Altogether, these findings indicate that degradation efficiency (as expressed by S_Deg) inversely correlates with phenol selectivity: catalysts that minimize the formation of CO₂ and ring-opening by-products are more effective in preserving phenol once formed.

This highlights the importance of engineering copper oxide catalysts not just for benzene activation, but also for suppressing non-selective oxidation pathways through careful control of precursor chemistry, calcination conditions, and surface redox properties.

2.1.2. Photocatalytic Degradation of Phenol Under UV Light Irradiation

To better elucidate the origin of the observed differences in phenol selectivity and yield among the catalysts, additional photocatalytic tests were conducted using phenol as the starting substrate under UV irradiation. These experiments aimed to directly evaluate the catalysts’ tendency to overoxidize phenol, which is a known limitation in benzene hydroxylation reactions. The results are summarized in

Table 3. Phenol degradation over all CuO-based photocatalysts under UV irradiation for 180 min.

Photocatalyst	Irradiation Time (min)	Phenol Degradation (%)
NCuO 10 min	180	~2
NCuO 2 h	180	~1
ACuO 10 min	180	~19
ACuO 2 h	180	~18
CuO	180	~3

.

The results reported in Table 3 provide important insights into the different behaviors of the catalysts regarding phenol stability under UV irradiation. Notably, NCuO 10 min and NCuO 2 h exhibit minimal phenol degradation (~2% and ~1%, respectively) after 180 min of irradiation. This behavior is fully consistent with the higher phenol selectivity and yield observed for these samples (see Figure 4 and Table 3), especially considering that phenol is an intermediate highly susceptible to overoxidation. The suppressed phenol degradation confirms that these nitrate-derived catalysts facilitate not only the selective hydroxylation of benzene but also inhibit further mineralization of the target product. In contrast, ACuO samples show significantly higher phenol degradation (18–19%), despite achieving lower phenol selectivity and yield under the same reaction conditions. This indicates that once formed, phenol is readily overoxidized on these catalysts, which is likely due to differences in surface properties or charge carrier dynamics associated with the precursor and calcination treatment. The commercial CuO also promotes partial phenol degradation (~3%) and shows the lowest selectivity (~3.2%), aligning with its tendency to favor complete oxidation pathways over partial hydroxylation.

These observations further reinforce the interpretation that NCuO catalysts, particularly NCuO 10 min, achieve a better balance between activity and selectivity, not only by promoting benzene hydroxylation but also by preserving phenol from overoxidation.

Overall, the phenol degradation experiments validate the trends observed in yield and selectivity and highlight the importance of tuning the catalyst composition to prevent phenol overoxidation while maintaining sufficient benzene conversion.

Generally, bare semiconductors such as TiO₂ and ZnO exhibit limited photocatalytic activity in the selective oxidation of benzene to phenol, typically yielding poor phenol yield and selectivity [70,71,72,73,74,75]. These limitations arise primarily from the rapid recombination of photogenerated charge carriers and the inherently non-selective action of hydroxyl radicals, which tend to promote overoxidation and the formation of undesired by-products such as CO₂ [76,77]. While ZnO offers better electron mobility than TiO₂, its photocatalytic stability is hindered by photocorrosion under UV irradiation, further limiting its performance [78,79]. Notably, the NCuO catalysts tested in this work far exceed the performance of bare CuO and conventional semiconductor oxides. Specifically, NCuO 2 h and NCuO 10 min achieve phenol yields of ~10.5% and 10.0%, respectively—nearly double that of bare CuO (2.8%). This enhancement highlights the strong potential of nitrate-derived CuO systems as efficient, selective, and practical photocatalysts for the hydroxylation of benzene under mild UV irradiation.

2.1.3. Mechanistic Investigation of NCuO (10 min) Through HPLC Monitoring

To gain mechanistic insights into the benzene hydroxylation process, product distribution was monitored over time using HPLC analysis, but only for the NCuO 10 min catalyst. This sample was selected due to its optimal balance of activity and phenol selectivity, making it the most representative candidate to investigate the reaction pathway. The resulting concentration profiles are reported in the Supporting Information (Figure S2), while the proposed reaction mechanism is illustrated in Scheme 1.

The analysis shows that phenol is the main oxidation product, forming rapidly within the first 60 min of irradiation and then progressively decreasing. This behavior reflects its role as a primary intermediate, which undergoes further oxidation under prolonged UV exposure. Among the by-products, catechol is formed through ortho-hydroxylation of phenol. Its concentration increases steadily and then stabilizes, suggesting that while it resists rapid degradation, it is eventually oxidized in slower steps—likely into ring-opening compounds or CO₂. Hydroquinone, the para-hydroxylated counterpart, appears in lower amounts, likely due to its fast transformation into p-benzoquinone. Notably, p-benzoquinone accumulates initially but subsequently decreases, indicating that it is not terminal, and is further oxidized, ultimately contributing to CO₂ formation. The absence of resorcinol implies that meta-hydroxylation is not a preferred pathway in this system.

These findings form the basis for the proposed reaction mechanism depicted in Scheme 1, which was developed directly from the HPLC time-course data. In this scheme, benzene is first converted to phenol, which subsequently undergoes oxidation to catechol and hydroquinone. The latter is then transformed into p-benzoquinone. Both catechol and p-benzoquinone eventually participate in further degradation pathways, ultimately leading to ring-opening products and CO₂.

2.2. Decomposition of Formic Acid to Hydrogen and Carbon Dioxide

The CuO-Cu₂O materials were further evaluated for the catalytic decomposition of formic acid into H₂ and CO₂. Batch experiments were conducted in round-bottom flasks held at the target reaction temperature, each equipped with a magnetic stir bar and a reflux condenser to prevent evaporation of solvent or FA, under a nitrogen atmosphere. Initial tests were performed in a stainless-steel pressure reactor (Entry 1,

Table 4. Decomposition of formic acid to hydrogen and carbon dioxide promoted by CuO–Cu₂O.

Entry a	Catalyst (g)	T (°C)	Time (h)	Conversion (%) b
1	NCuO 10 min (0.1)	100	1	>99
2	NcuO 2 h (0.1)	100	1	>99
3	AcuO 10 min (0.1)	100	1	>99
4	AcuO 2 h (0.1)	100	1	>99
5	NcuO 10 min (0.1)	25	1	>99
6	NcuO 2 h (0.1)	25	1	>99
7	AcuO 10 min (0.1)	25	1	>99
8	AcuO 2 h (0.1)	25	1	>99
9	NcuO 10 min (0.05)	25	0.25	>99
10	NcuO 2 h (0.05)	25	0.25	>99
11	AcuO 10 min (0.05)	25	0.25	>99
12	AcuO 2 h (0.05)	25	0.25	>99

^a All reactions were carried out in 4.0 mL of water as solvent, [HCOOH]₀ = 1.3 M, N₂ atmosphere. ^b Determined by ¹H NMR spectroscopy (CD₃OD, 298 K) using 1,3,5-trimethoxybenzene as internal standard.

); the gaseous effluent was analyzed by gas chromatography to confirm exclusive H₂ and CO₂ formation and to verify the complete suppression of CO production (see Figure S3). Under the same reaction conditions, no H₂ formation was observed when using commercial CuO.

The influence of reaction temperature on catalytic performance was initially investigated, yet no appreciable differences in activity were observed among the four catalysts evaluated (entries 1–4, Table 4). All the catalysts afforded FA conversion under these reaction conditions. Consequently, the reaction temperature was lowered, observing again quantitative conversions (entries 5–8, Table 4). Owing to the exceptionally high conversions even at ambient temperature, subsequent experiments employed shortened reaction times and reduced catalyst loadings; despite these more stringent conditions, formic acid was still converted quantitatively (entries 9–12, Table 4). These preliminary findings demonstrate that the CuO-Cu₂O systems exhibit superior activity relative to previously reported catalysts, which typically require temperatures above 200 °C to reach comparable conversions [48,49,50]. Moreover, given the high FA conversions achieved under mild reaction conditions and in the absence of added bases commonly employed to facilitate dehydrogenation [80,81]. It will be worthwhile in future work to support the CuO–Cu₂O catalysts on suitable carriers to reduce the overall copper loading.

3. Materials and Methods

3.1. Materials

All chemicals were of analytical grade and used without further purification. Copper(II) nitrate hydrate (Cu(NO₃)₂·3H₂O ≥ 99.9%, Sigma-Aldrich, St. Louis, MO, USA) and Copper(II) acetate monohydrate (CH₃COO)₂Cu·H₂O ≥ 99.9%, Sigma Aldrich, St. Louis, MO, USA) were used as precursors of the (photo)catalyst. Commercial CuO (>99%, Sigma Aldrich, St. Louis, MO, USA) was used as a benchmark sample for comparison with the synthesized catalysts in terms of structural, morphological, and photocatalytic properties. Benzene (≥99%, Sigma Aldrich, St. Louis, MO, USA) was used as the target compound for photocatalytic oxidation. Acetonitrile (≥99.9%, Sigma Aldrich, St. Louis, MO, USA) was employed as a co-solvent. Hydrogen peroxide (30% w/w water solution) served as the oxidizing agent. Distilled water was used in all solution preparations.

3.2. Catalysts Preparation

The catalysts were synthesized using two different precursors: copper acetate and copper nitrate. For the synthesis with copper acetate, 2 g of copper(II) acetate monohydrate were dissolved in 50 mL of distilled water. Subsequently, an ammonia solution (30% w/w) was added dropwise to adjust the pH to 11 using a pH meter (Mettler Toledo FE20-Basic Five Easy™, Sigma Aldrich, St. Louis, MO, USA). The resulting solution was subjected to sonication at room temperature for 3 h, during which the pH was continuously monitored [82]. The resulting precipitate was collected by centrifugation, washed several times with distilled water, and then dried at 90 °C for approximately 12 h. Finally, the dried material was calcined in a muffle furnace at 500 °C, with a heating rate of 5 °C/min. Two different calcination times were applied: 10 min and 2 h, in order to evaluate the effect of thermal treatment duration on the structural and (photo)catalytic properties of the final material. The series of copper oxide catalysts synthesized in this study are designated as NCuO 10 min and NCuO 2 h for those derived from copper nitrate, and ACuO 10 min and ACuO 2 h for those derived from copper acetate, with the suffix indicating the calcination time.

3.3. Chemical-Physical Characterization Methods

Wide-angle X-ray diffraction (WAXD) patterns were obtained using an automatic Bruker D8 Advance diffractometer in reflection geometry (Bruker, Billerica, MA, USA) equipped with nickel-filtered Cu-Kα radiation (λ = 1.5406 Å). The average crystallite size (D) was calculated using the Scherrer equation [83]:

$$D={\displaystyle \frac{K\lambda }{\beta \mathit{cos}\theta }}$$

(1)

where K is the shape factor (taken as 0.89), λ is the X-ray wavelength, β is the full width at half maximum (FWHM) of the selected diffraction peak (in radians), and θ is the Bragg angle. Laser Raman spectra were obtained at room temperature with a Dispersive MicroRaman (Invia, Renishaw, Wotton-under-Edge, UK), equipped with a 514 nm laser, in the range 100–900 cm⁻¹ Raman shift. The Brunauer, Emmett, and Teller (BET) surface area of the samples was measured from dynamic N₂ adsorption measurement at −196 °C, performed by a Costech Sorptometer 1042 (Costech International S.p.A., Milan, Italy); all samples before the measurement were pretreated at 150 °C for 30 min in He flow. FTIR adsorption spectra were carried out using a Bruker Vertex 70 spectrometer (Bruker, Billerica, MA, USA). Spectra were recorded directly on composites, with 32 scans collected in the range of 450–700 cm⁻¹ at a resolution of 2 cm⁻¹.

3.4. Photocatalytic Tests

An aqueous solution of benzene was prepared at a concentration of 25.6 mmol/L in distilled water. To this, 2.3 mL of acetonitrile was added to improve the solubility of benzene. The resulting mixture was sonicated for 20 min to ensure complete homogenization. The solution (50 mL total volume) was then transferred into a cylindrical batch-type photoreactor made of Pyrex, equipped for continuous stirring and temperature control via a water-cooling system. The photocatalyst (5 mg) was added to the solution. The suspension was left under stirring in the dark for 30 min to allow adsorption equilibrium between benzene and the catalyst surface. After the dark adsorption step, 0.56 mL of hydrogen peroxide was added to the reactor as the oxidant. The reaction mixture was then irradiated uniformly using a 12 W UV LED strip light (BuyLEDStrip.com, Denethor 17, 5663RL, Geldrop, The Netherlands).

The photocatalytic reaction was monitored over a total duration of 7 h. Samples were withdrawn at the following time intervals: 15, 30, 45 min; 1, 2, 3, 4, 5, 6, and 7 h. Two types of analyses were conducted for each time point: gas chromatography (GC) and high-performance liquid chromatography (HPLC).

For GC analysis, 0.20 mL of reactor solution was collected using a graduated syringe and diluted in a vial with 0.80 mL of acetonitrile.

For HPLC analysis, 0.50 mL of the reactor solution was diluted with 0.50 mL of distilled water. GC was used to determine the concentration and conversion of benzene. Analyses were carried out using a capillary column (DB Heavy Wax, 30 m × 0.35 mm i.d. × 0.25 µm film thickness, Thermo Fischer Scientific, Waltham, MA, USA). The carrier gas was helium at a constant flow rate of 1 mL/min. The column temperature was programmed as follows: 40 °C for 2 min, ramped at 5 °C/min to 90 °C, then 20 °C/min to 250 °C and held for 10 min. The injector operated in split mode (10:1) at 189 °C, and the detector temperature was set to 300 °C.

HPLC was employed to quantify phenol and other aromatic oxidation products, including hydroquinone, catechol, resorcinol, and 1,4-benzoquinone. A Dionex UltiMate 3000 (Thermo Fischer Scientific, Waltham, MA, USA) system equipped with a DAD detector, column oven, and automatic injector with a 100 µL sample loop was used.

Chromatographic separation was achieved using a Phenomenex Luna C-18 (Thermo Fischer Scientific, Waltham, MA, USA) column (150 × 4.6 mm i.d.; 5 µm particle size) at 35 °C. The mobile phase consisted of water (solvent A) and acetonitrile (solvent B), with the following elution program: 0–14 min: 15% B (isocratic); 14–23 min: linear gradient from 60% to 100% B, 23–30 min: 100% B (isocratic).

The flow rate was set at 0.8 mL/min, and the injection volume was 50 µL. Detection was carried out at 270 nm.

The following mathematical formulas for the determination of benzene conversion, phenol yield, selectivity to phenol, and selectivity to by-products were used:

$$benzene conversion =\left(1-{\displaystyle \frac{C}{C_0}}\right)\times 100$$

(2)

$$yield to Phenol=\left({\displaystyle \frac{S }{Q_0}}\right)\times 100$$

(3)

$$selectivity to P=\left({\displaystyle \frac{S (or Deg)}{S_{TOT }+Deg}}\right)\times 100$$

(4)

where

C₀ = benzene concentration after the dark period (mmol/L);

C = benzene concentration at the generic irradiation time (mmol/L);

Q₀ = moles of benzene in solution after the dark period (mmol);

P = reaction product (phenol or hydroquinone or catechol or resorcinol or p-benzoquinone);

S_TOT = total moles of the reaction products in the liquid phase detected by HPLC (mmol);

S = moles of phenol in liquid phase (mmol);

Deg = ring-opening products and CO₂ = (benzene reacted−S_TOT).

3.5. Catalytic Tests for Formic Acid Decomposition

Nuclear magnetic resonance (NMR) spectra were acquired on Bruker Avance spectrometers (Bruker, Billerica, MA, USA) operating at 400 MHz and 300 MHz for ¹H NMR. Chemical shifts (δ, ppm) are reported relative to tetramethylsilane as an external standard, with the residual protio signals of the deuterated solvents used as internal references. H₂ and CO₂ were identified with a Trace1300 gas chromatograph coupled with a thermal conductivity detector (GC-TCD) and a TG-BOND Q column (30 m × 0.32 mm × 10 μm from Thermo Scientific, USA). The formic acid decomposition tests were carried out following a procedure reported in the literature. Referring to Entry 1 in Table 3, a 25 mL round-bottomed flask equipped with a magnetic stir bar and a refrigerant was charged with formic acid (0.234 g, 5 mmol), water (4 mL), and catalyst (0.1 g). The mixture was stirred for 1 h at 100 °C. The flask was cooled at room temperature and a solution of 1,3,5-trimethoxybenzene (0.093 g, 0.5 mmol, in 4 mL of methanol) was added as an internal standard. The catalyst was separated by centrifugation, and the filtrate was analyzed by ¹H NMR spectroscopy using methanol-d₄ as a solvent, determining a yield of >99% (see Figure S4 in the Supporting Information). Calibrating the signal relative to the peak of the standard at 100, the following equation was used to calculate the formic acid (FA) conversion:

Conversion (%) = 100 − FA integral value

(5)

The same reaction was also performed in a stainless-steel pressure reactor under the same conditions. After 1 h, 100 µL from the gas phase reaction mixture in the reactor was injected with a Hamilton^®Gastight^® syringe (Sigma-Aldrich, St. Louis, MO, USA) and analyzed by GC-TCD, observing the only presence of H₂ and CO₂ (see Figure S3 in the Supporting Information).

4. Conclusions

This work explored the photocatalytic hydroxylation of benzene to phenol and the thermal decomposition of formic acid using CuO-based catalysts synthesized from copper nitrate (NCuO) and copper acetate (ACuO) precursors, with calcination times of 10 min and 2 h. Structural and spectroscopic characterizations (WAXD, Raman, and BET) revealed differences in crystallinity, surface area, and Cu(I)/Cu(II) speciation depending on the precursor and thermal treatment. Raman analysis, in particular, suggested the formation of minor Cu₂O domains in NCuO samples, which may play a role in modulating photocatalytic behavior.

Photocatalytic tests for benzene hydroxylation demonstrated that all synthesized catalysts outperformed commercial CuO in terms of phenol yield and selectivity. Among them, the nitrate-derived samples, especially NCuO 10 min, showed the best balance between benzene conversion and phenol selectivity, achieving a phenol yield of ~10% and selectivity up to 19% within just 60 min of UV irradiation. In contrast, ACuO and commercial CuO samples led to higher overall benzene conversion but suffered from rapid phenol degradation and lower selectivity. Complementary phenol degradation tests confirmed that NCuO catalysts exhibit reduced overoxidation activity, likely due to improved phenol desorption and moderated radical reactivity.

Additionally, all CuO-based catalysts deliver quantitative formic acid conversion to H₂ and CO₂ with complete suppression of CO formation across a broad temperature range, including ambient conditions, while maintaining high activity under shortened reaction times and minimal catalyst loadings. This performance surpasses that of previously reported systems, which typically require temperatures above 200 °C, underscoring the potential of the tested catalysts as a cost-effective, base-free catalyst for formic acid dehydrogenation.

Taken together, these findings underscore the versatility of CuO-based catalysts derived from inexpensive, earth-abundant precursors. The results highlight the crucial influence of precursor chemistry and calcination time on catalytic performance and selectivity, both in photocatalytic benzene hydroxylation and in low-temperature hydrogen production.

Future work will focus on immobilizing these materials on high-surface-area supports to enhance dispersion and reduce copper usage while maintaining catalytic performance. In parallel, more advanced spectroscopic investigations, such as XPS, will be necessary to elucidate the defective structure of NCuO better and confirm the nature and distribution of Cu(I) domains potentially responsible for the observed selectivity.