Synthesis of MOF-Derived Mono-, Bi- and Trimetallic Fe, Zn and Cu Oxides for Microwave-Assisted Benzyl Alcohol Oxidation

Abstract

The increasing demand for sustainable chemical processes has fostered the development of advanced catalytic systems for biomass valorization. In this work, a series of mono-, bi-, and trimetallic oxides (FeO, FeCuO, FeZnO, and FeCuZnO) were successfully synthesized using MIL-101-based MOFs as sacrificial templates. The obtained materials were thoroughly characterized by N₂ adsorption–desorption, XRD, FTIR, and TEM/STEM-EDX to investigate their structural, morphological, and textural properties. Their catalytic performance was evaluated in the selective oxidation of benzyl alcohol, a lignin-derived platform molecule, into benzaldehyde under microwave irradiation as a sustainable heating strategy. The results demonstrate that MOF-derived oxides exhibit superior activity compared to their parent MOFs, highlighting the beneficial effect of thermal treatment on the exposure of active sites. Among the catalysts, heterometallic oxides showed enhanced performance due to synergistic effects between metals. In particular, FeZnO reached a maximum yield of 62.1% towards benzaldehyde at 150 °C and 30 min, outperforming the monometallic oxide. Recycling tests revealed that FeZnO retained higher overall performance than FeCuO, which suffered from progressive copper leaching. These findings confirm the potential of MOF-derived multimetallic oxides as efficient and reusable heterogeneous catalysts for selective biomass-derived alcohol oxidation. The combination of microwave-assisted processes and the tuneable nature of MOF-derived oxides provides a promising pathway for designing sustainable catalytic systems with industrial relevance.

1. Introduction

The contemporary era has been characterised by significant population growth and urbanisation, which have led to a sharp increase in both energy demand and chemical production as compared to previous decades. To meet these demands, the chemical industry has traditionally relied on fossil fuels such as oil, coal and natural gas. However, these non-renewable resources are responsible for more than 75% of global greenhouse gas emissions, thus, contributing substantially to climate change and environmental degradation [1]. This has intensified the search for sustainable alternatives.

Among the different options, lignocellulosic biomass has attracted considerable attention due to its abundance, global availability, and potential to be transformed into a wide range of valuable products [2,3]. Biomass is regarded as carbon-neutral, since the CO₂ released during its transformation is reabsorbed by photosynthetic organisms, enabling a short carbon cycle [4,5]. Importantly, lignocellulosic biomass does not compete with food production, making it one of the most promising feedstocks for sustainable chemical processes. It is mainly composed of cellulose (40–50%), hemicellulose (20–30%), and lignin (20–25%), with minor contributions from proteins, lipids, soluble sugars, and minerals [5,6].

While cellulose and hemicellulose are carbohydrate polymers, lignin is a complex hydrophobic polymer formed from p-coumaryl, coniferyl, and sinapyl alcohols. Its predominant linkage, the β-O-4 bond, represents 60–70% of its structure [7,8,9]. Controlled depolymerisation of lignin enables the release of simple aromatic molecules that can serve as platform compounds [8]. Among these, benzyl alcohol is especially relevant, since it can be selectively oxidised to benzaldehyde, an intermediate of broad industrial application in perfumes, dyes, pharmaceuticals, and agrochemicals [10].

Traditionally, benzaldehyde has been obtained through hydrolysis of benzyl chloride [11] or direct oxidation of toluene [12,13]. However, these processes present severe drawbacks, including toxic by-products and contamination of the final product. As a result, heterogeneous catalysis has emerged as a more sustainable strategy, enabling selective oxidations under milder conditions with reduced environmental impact [5].

MOFs are coordination polymers in which metal centres are connected by organic ligands to form ordered one-, two-, or three-dimensional structures [3,14]. Their exceptional crystallinity, porosity, and tuneable surface properties make them versatile for a wide range of applications, including gas storage, separation, sensing, drug delivery, and catalysis [3,14,15,16,17,18,19,20,21].

Nevertheless, pristine MOFs often suffer from limited chemical and thermal stability, which restrict their direct use in catalysis. To address this limitation, MOFs are increasingly employed as sacrificial templates to generate more robust functional materials, such as metal oxides, metal nanoparticles embedded in carbon matrices, or single-atom catalysts (SACs) [3,22,23]. These MOF-derived materials combine the morphological advantages of the parent frameworks with improved durability, while offering higher surface areas, more accessible active sites, and enhanced catalytic stability [24,25,26].

Therefore, MOF-derived materials lead to the formation of their respective metal oxides upon carbonisation. Historically, metal oxides have served as excellent heterogeneous catalysts in a wide variety of reactions, as summarised in

Table 1. Oxide catalysts in heterogeneous catalysis.

Catalyst	Reaction	Optimum Reaction Condition			Yield (%)	Ref.
		Temp (°C)	Time (h)	CL (%)
CFA-ZnO	Biodiesel synthesis	140	3	0.5	83.17	[27]
WZC (WO3-Zr2O3/CaO)	Esterification and transesterification	80	1	2	94.1	[28]
SrO–ZnO/Al2O3	Transesterification	70	3	10	95.1	[29]
Fe2O3MnOSO42−/ZrO2	Transesterification of glycerol	180	4	3	96.5	[30]
RuNi-CNTs-X	Hydrogenation of benzene	50	6	0.4	70.9	[31]

.

In this context, the synthesis of MOF-derived metal oxides provides a promising route to fabricate efficient catalysts for selective oxidation. During thermal treatment, the organic ligands decompose, leading to the formation of oxides with high specific surface areas and increased numbers of active centres. Such materials frequently outperform their MOF precursors due to improved stability and catalytic efficiency [23,25,26].

In this work, MOFs and their corresponding oxides were synthesised and evaluated as catalysts for the selective oxidation of benzyl alcohol into benzaldehyde under microwave-assisted conditions. Microwave irradiation was chosen as a sustainable alternative to conventional heating, since it allows achieving comparable results in significantly shorter times owing to its homogeneous and efficient energy transfer [32]. In line with the principles of green chemistry, we focused on earth-abundant and low-cost metals, avoiding the use of noble metals such as Pt, Pd, Au, Ag, Ru, or Rh, which are traditionally applied in alcohols oxidation [33,34,35]. Iron-based catalysts were selected because Fe is inexpensive, abundant, and able to generate reactive oxygen species (ROS) in the presence of oxidants, thereby promoting selective oxidation [36].

To further enhance catalytic efficiency, recent studies have emphasised the importance of synergistic interactions between multiple metals. Incorporating two or more metals into a catalyst can improve redox behaviour, surface acidity, and stability. In this study, bi- and trimetallic MOFs were synthesised and transformed into the corresponding oxides through calcination, exploiting these synergistic effects [37,38]. Notably, Fe has demonstrated cooperative behaviour when combined with Cu [39] or Zn [40], suggesting that Fe–Cu–Zn systems could offer superior catalytic performance.

To contextualise the relevance of this research,

Table 2. Catalysts and reactions carried out for the selective oxidation of benzyl alcohol.

Catalyst	Time (h) and Temperature (°C)	Oxidant	Yield (%)	Ref
Pd/PILs	110 °C, 4 h	4 MPa O2	82	[41]
Fe: Pd NPs	4 h	1 bar of O2	86	[42]
Cu–Cr-HT	210 °C, 5 h	O2 (6 mL/min)	35	[43]
Ti(acac 1)-SBA-15	100 °C, 7 h	TBHP (tert-butyl hydroperoxide)	90	[44]
NiFe2O4 NPs	60 °C, 3 h	TBHP	85	[45]

¹ acac: acetylacetonate.

summarises representative studies from the literature on the selective oxidation of benzyl alcohol to benzaldehyde, highlighting catalysts, reaction conditions, and yields achieved.

Therefore, the aim of this research was to synthesise and characterise mono-(Fe), bi-(Fe–Cu, Fe–Zn), and trimetallic (Fe–Cu–Zn) MOFs and their derived oxides, and to evaluate their performance in the microwave-assisted selective oxidation of benzyl alcohol. Special emphasis was placed on verifying that oxide derivatives outperform their respective MOF precursors, in agreement with previously reported in the literature, and on elucidating the role of synergistic interactions in multimetallic systems.

2. Results and Discussion

2.1. Characterization Results

2.1.1. N₂ Physisorption

Figure 1 shows the N₂ adsorption–desorption isotherms (−196 °C) of the synthesised MOFs and their corresponding derived oxides. In each case, the isotherms of the parent MOF (light trace) and the calcined oxide (dark trace) are compared, grouped according to metal composition. According to the IUPAC classification, MOFs usually display type I isotherms, characteristic of microporous materials. However, in all cases studied, type IV isotherms were observed for both MOFs and their calcined derivatives, indicative of mesoporous materials’ behaviour.

As shown in Figure 1, the metal oxides adsorb significantly larger N₂ volumes than their MOF precursors, most likely due to enhanced pore accessibility after thermal treatment. This phenomenon can be attributed to the removal of organic linkers and the concomitant formation of interparticle porosity [46].

In addition, the bimetallic and trimetallic oxides (FeCuO, FeZnO, FeCuZnO) exhibit isotherms with similar profiles and higher N₂ volume adsorbed as compared to the monometallic oxide (FeO). This suggests that the incorporation of more than one metal contributes to an increased pore volume.

Table 3. Values obtained from adsorption isotherm range of surface area, total pore volume and mesopore volume.

Catalyst	SBET (m2/g)	VBJH (cm3/g)	VMESO (cm3/g)
Fe-MOF	6.9	0.022	0.009
FeCu-MOF	19.3	0.036	0.012
FeZn-MOF	24.4	0.062	0.023
FeCuZn-MOF	6.1	0.012	0.008
FeO	11.6	0.029	0.014
FeCuO	16.4	0.091	0.018
FeZnO	19.7	0.094	0.024
FeCuZnO	16.3	0.072	0.019

summarises the specific surface area, total pore volume, and mesopore volume for all the samples, calculated from the N₂ adsorption–desorption isotherms of the synthesised materials.

As anticipated from the isotherms shown in Figure 1, the surface area values obtained for the MOFs are significantly lower than those reported in the literature for MIL-101-type structures [47]. In contrast, the derived metal oxides exhibit higher surface areas than their respective MOF precursors. Unexpectedly, this behaviour is opposite to the expected decrease in surface area typically associated with framework collapse upon calcination. Such a discrepancy can be rationalised by considering the synthesis conditions described previously: specifically, the substitution of hydrofluoric acid (HF) with hydrochloric acid (HCl) may have affected the integrity and porosity of the material. While HF promotes the formation of well-defined, highly porous structures due to its strong nucleation capacity, HCl lacks the same efficiency [48].

A general increase in both total and mesopore volume is also observed after thermal treatment and transformation of the MOFs into oxides. This effect is particularly pronounced in the bimetallic and trimetallic oxides and can be attributed to the generation of interparticle porosity upon removal of the organic linker.

2.1.2. X-Ray Diffraction

Figure 2 depicts the diffraction patterns of MOFs (Figure 2a) and oxides derived from them (Figure 2b).

Figure 2 shows the XRD patterns of the synthesised MOF-type materials. The diffraction peaks observed match the characteristic crystallographic pattern of Fe-MIL-101 reported in the literature, with intense reflections at 2θ = 9.0°, 9.5°, 12.5°, and 18.4° [49,50].

Upon incorporation of a second metal, such as copper (FeCu-MIL-101), new and more intense peaks appear at 2θ = 8.8°, 12.2°, 17.2°, 19.9°, and 25.0°. This behaviour can be attributed to competition between Cu and Fe for coordination with the organic ligand, resulting in structural modifications in which the crystalline contribution of the Cu-MOF becomes dominant [49]. A similar trend is observed for the other bimetallic MOF (FeZn-MIL-101), where the main reflections (2θ = 9.0°, 10.3°, and 12.1°) correspond to the Zn-MOF structure. However, in this case the signal intensity is lower and the peaks are broader, suggesting a more amorphous character. This loss of structural order may arise from two phenomena: (i) the “breathing effect,” in which the framework expands or contracts due to hydrogen-bonding interactions with the solvent (DMF), causing significant variations in the unit cell; and (ii) crystalline disorder associated with the simultaneous incorporation of metal ions of different ionic ratio [51,52]. Finally, the trimetallic FeCuZn-MIL-101 displays an intermediate diffraction pattern between those of the bimetallic phases, exhibiting greater disorder than FeCu-MIL-101 but more structural organisation than FeZn-MIL-101. This behaviour may be linked to distortions induced by the simultaneous incorporation of three cations; in particular, the reduced crystallinity appears to be mainly associated with Zn, consistent with the behaviour previously observed.

Overall, the results suggest that the incorporation of more than one metal does not lead to the formation of a new crystalline phase, but rather to the coexistence of phases corresponding to the individual monometallic MOFs, which appear to be well integrated within the same material as happened in other similar studies [15].

Figure 2b presents the XRD patterns of the MOF-derived oxides. All samples display a defined crystalline structure, with reflections at 2θ = 24.0°, 33.0°, 36.0°, 41.0°, 49.0°, 54.0°, 58.0°, 62.0°, and 64.0°, corresponding to the (012), (104), (110), (113), (024), (116), (018), (214), and (300) planes, respectively [53]. These reflections are characteristic of hematite (α-Fe₂O₃), indicating that the dominant crystalline phase in all oxides is determined by the calcination of the Fe-MOF precursor. In contrast, Cu and Zn oxides appear to be structurally amorphous, since no characteristic diffraction peaks of these compounds were detected.

2.1.3. FTIR-ATR

Attenuated total reflectance Fourier-transform infrared spectroscopy (FTIR-ATR) was employed to investigate the surface structure of the synthesised materials. The corresponding spectra are shown in Figure 3.

Figure 3a shows the FTIR-ATR spectra of the synthesised MOF-type materials. In all cases, characteristic bands associated with MIL-101-type structures are observed, consistent with the presence of the terephthalic acid linker used in the synthesis. Specifically, the bands at 750 and 1015 cm⁻¹ correspond to C–H bending vibrations, while the band at 1505 cm⁻¹ confirms the presence of C=C stretching modes of the aromatic ring. In addition, the bands at 1400 and 1600 cm⁻¹ are attributed to the symmetric and asymmetric stretching vibrations, respectively, of the carboxylate group (–COO⁻). The band at 1663 cm⁻¹ is related to C=O stretching of uncoordinated carboxyl groups [45]. Interestingly, for the Zn-containing materials (FeZn-MIL-101 and FeCuZn-MIL-101), an additional band at 1750 cm⁻¹ is detected, which can be assigned to the C=O stretching of free carboxylic acid groups (COOH), possibly indicating the presence of non-coordinated groups or residual solvent interactions. Finally, weak bands at around 600 cm⁻¹ are associated with metal–oxygen (M–O) vibrations [15,54]. Overall, these results confirm a similar organic surface structure for all MOFs, in line with the use of the same organic linker in their synthesis, and are consistent with the structural features of MIL-101-type frameworks.

For the thermally treated materials, shown in Figure 3b, no bands corresponding to the organic groups of H₂BDC are observed, confirming their removal after high-temperature calcination. A weak band at approximately 1740 cm⁻¹ is still present, associated with residual C=O vibrations, which may arise from incomplete decomposition of the linker. This suggests that higher calcination temperatures or longer treatment times could be required to ensure total removal of organic moieties. Finally, all derived oxides exhibit a band around 600 cm⁻¹, corresponding to M–O stretching vibrations [15].

2.1.4. HR-TEM

To investigate the particle morphology of the synthesised materials, the bimetallic and trimetallic oxides were analysed by high-resolution transmission electron microscopy (HR-TEM). The corresponding images are shown in Figure 4.

For the bimetallic FeCuO oxide (Figure 4a), well-defined cylindrical structures with large diameters, regular edges, and compact morphology can be distinguished. The magnified image (bottom right inset) reveals a rough surface, attributed to the loss of organic ligands during calcination, which results in the formation of superficial porosity [15].

In contrast, the incorporation of zinc in FeZnO (Figure 4b) leads to the development of two clearly differentiated structures. In the regions where iron is predominant (highlighted in blue), dense and irregular aggregates are observed. In the Zn-rich regions (marked in fuchsia and enlarged in the lower-right inset), smaller and more dispersed iron particles were detected, which may be related to a dispersion effect promoted by Zn during oxide formation.

Finally, the trimetallic FeCuZnO oxide (Figure 4c) again exhibits thick cylindrical structures, similar to those of FeCuO. This suggests that copper acts as the dominant structural element, favouring the formation of a more homogeneous and compact network.

2.1.5. STEM-EDX

STEM-EDX analysis was performed on the same materials studied by HR-TEM in order to evaluate the spatial distribution of metals in the catalysts. The technique was applied to the FeCuO, FeZnO, and FeCuZnO oxides to assess the dispersion of the metallic elements within each sample. The complete elemental mapping (C, O, Fe, Cu, and Zn) is provided in Figure A1 in Appendix A.

For FeCuO (Figure 5a), iron is predominantly distributed across the entire surface of the material, with copper present to a lesser extent. This observation suggests that Fe is the dominant element both structurally (as confirmed by XRD) and at the surface level in the final material.

In contrast, for FeZnO (Figure 5b), zinc is mainly located on the outer regions of the catalyst, while iron appears more confined in the interior. This result indicates that, despite following a similar synthetic protocol, the metals distribution in the overall material depends on their intrinsic nature and structural affinity, reflecting differences in their behaviour during the oxide formation.

For the trimetallic FeCuZnO oxide (Figure 5c), a mixed distribution of the three metals is observed. Certain regions resemble the distribution found in FeZnO, whereas in the majority of the material the configuration is closer to that of FeCuO. These analyses together with those obtained by HR-TEM suggests that copper plays a dominant role in controlling the structural and morphological organisation of the system, as its presence is directly associated with the cylindrical structures and the prevailing metallic distribution when the three metals are combined.

2.2. Catalytic Results

2.2.1. Comparation Between MOFs and Derived Metallic Oxides

The evaluation of the catalytic performance of the as-synthesised MOFs materials and their corresponding metal oxides was carried out using the selective benzyl alcohol oxidation to benzaldehyde. Typically, catalytic tests were carried out under the following standard conditions: 120 °C, 30 min, 0.1 M benzyl alcohol in acetonitrile, and a substrate-to-oxidant molar ratio of 1:4. In addition, control experiments were performed in the absence of both catalyst and oxidant, as well as in the absence of either one individually. In all cases, negligible conversion was observed; therefore, these results are not included in the figures. Thus, these controls highlight the essential role of the catalyst and of TBHP as oxidant, since its absence or use in insufficient proportions (substrate/oxidant < 1:4) resulted in nearly null conversion and selectivity.

The results related to the catalytic activity of the evaluated material in the microwave-assisted benzyl alcohol oxidation reaction are summarised in Figure 6, which shows, for each catalyst, the conversion of benzyl alcohol (blue line) together with the selectivity to benzaldehyde and benzoic acid (stacked bars).

The catalysts screening showed higher conversion values for the metal oxides (Figure 6b) as compared to their respective MOF precursors (Figure 6a), confirming that MOF-derived oxides exhibit superior catalytic activity, in agreement with previous reports [24]. This enhancement can be attributed to the removal of the organic linker during calcination, which increases the accessibility of exposed metallic active sites. The lower performance achieved by the MOF materials may also be related to their reduced crystallinity and specific surface area, as evidenced by the XRD patterns (Figure 2) and N₂ adsorption–desorption measurements (Table 2). These structural and textural limitations result in reduced access to active sites and, consequently, lower catalytic activity.

It is also noteworthy that the oxides display higher selectivity towards the desired product, benzaldehyde, whereas the MOFs tend to over-oxidise the substrate to benzoic acid. This behaviour may be associated with the stronger oxidative character of the MOFs, which is detrimental to selectivity in this reaction, thereby positioning the oxides as more promising candidates for optimising the catalytic conditions.

Within the group of metal oxides (Figure 6b), the bimetallic and trimetallic catalysts (FeCuO, FeZnO, FeCuZnO) exhibit clearly higher conversions than the monometallic FeO, evidencing a positive synergistic effect between the combined metals. These results reinforce the hypothesis that the incorporation of a second or third metal enhances catalytic activity [37,39]. Based on these findings, and given the superior performance of the oxides compared to MOF materials, the rest of this study focused exclusively on the comparative analysis and optimisation of the reaction conditions for the MOF-derived oxides.

2.2.2. Study of the Effect of Temperature on Conversion and Product Selectivity

To evaluate whether the initially selected temperature (120 °C) was optimal for the catalytic process, and to analyse the response of the catalytic system to this parameter, additional experiments were performed at 90 and 150 °C while keeping the other experimental conditions constant (30 min, 0.1 M benzyl alcohol, AB/TBHP molar ratio 1:4). The results are presented in Figure 7 (a: 90 °C, b: 120 °C, c: 150 °C).

The results show that the lowest temperature evaluates, at 90 °C, the catalysts exhibited a negligible catalytic activity. At 120 °C, however, a clear improvement is observed, with significantly higher catalytic performance. A further temperature raise up to 150 °C produces only a slight additional rise in conversion, indicating that 120 °C is already close to the optimal temperature for this system.

In terms of selectivity, the copper-containing oxides (FeCuO and FeCuZnO) display a stronger oxidative character, which results in a higher proportion of benzoic acid at elevated temperatures. Although STEM analysis revealed that copper is not predominantly located at the surface, the relatively high conversion values obtained for FeCuO suggest the presence of accessible active sites, or alternatively, that metal dispersion within the oxide matrix contributes to its catalytic activity.

By contrast, zinc-containing oxides show improved selectivity towards benzaldehyde. According to STEM, Zn is preferentially located at the surface, which could enhance the reactivity of the system and favour partial oxidation. In addition, the Zn-based oxides exhibited the largest surface area and porosity (Table 2), which would promote efficient diffusion of reactants to active sites and thereby explain their catalytic behaviour. Moreover, according to the literature, ZnO not only exhibits semiconducting properties but also possesses Lewis acidity, which could be directly involved in the reaction mechanism or activate a secondary pathways [55]. In addition, the synergistic effect between Zn and other transition metals, an interaction that as has been demonstrated in previous studies may have a positive impact on the selectivity towards benzaldehyde [56]. Both hypotheses are plausible and open new perspectives for further investigation, which should consider: (i) the influence of the Zn content, (ii) the intrinsic effect of Zn, and (iii) the use of additional characterization techniques to correlate these results with the physicochemical properties of the catalysts.

Although benzoic acid was detected as a by-product under some conditions, benzaldehyde remained the major product in all cases. These results confirm the potential of the synthesised materials as promising catalysts for selective oxidation processes.

2.2.3. Study of the Effect of Reaction Time on Conversion and Product Selectivity

To investigate the kinetic evolution of the oxidation reaction and optimise the operating time, FeCuO was selected as catalyst, since it showed the best compromise between high conversion and limited by-product formation in the previous tests. The reaction was conducted at 150 °C, and different reaction times (5, 15, 30, and 45 min) were evaluated while keeping the other experimental parameters constant (0.1 M benzyl alcohol, AB/TBHP molar ratio 1:4).

As shown in Figure 8, the conversion of benzyl alcohol increases progressively with reaction time, reaching a maximum of 71.2% at 45 min. This behaviour can be attributed to the longer contact time between the substrate and the catalyst, which allows more reactant molecules to access the active sites.

However, the increase in conversion is accompanied by a loss of selectivity towards benzaldehyde, particularly evident at 45 min, where a significant amount of benzoic acid is formed. This effect can be explained by the consecutive oxidation of benzaldehyde: as it remains longer in the system and adsorbed on the catalyst surface, it is further oxidised to benzoic acid, thereby reducing the proportion of the targeted product and the overall yield.

Based on these results, 30 min has been selected as the optimal reaction time, as it provides a suitable balance between conversion (69.7%) and selectivity to benzaldehyde (81.2%). Under these conditions, a final yield of 56.5% is achieved, compared to only 35.9% at 45 min.

2.2.4. Recyclability Study

Finally, the reusability of the bimetallic catalysts that exhibited the highest yields—FeCuO (56.5%) and FeZnO (62.1%)—was evaluated under the previously optimised reaction conditions (150 °C, 30 min, BA/TBHP molar ratio 1:4), as shown in Figure 7c. Between successive cycles, the catalysts were subjected to a cleaning protocol consisting of recovery by centrifugation, followed by two washing steps with acetonitrile and two with acetone. The solids were then dried at 80 °C to remove residual solvent before being used in the next catalytic cycle. The results corresponding to the recyclability study are presented on Figure 9.

The results show that, for both catalysts, the conversion of benzyl alcohol to benzaldehyde continuously decreases with successive reuse cycles. In the case of FeCuO (Figure 9a), the conversion drops from 69% in the first run to 63% in the second, while for FeZnO (Figure 9b) it decreases from ~62% to 46% under the same conditions. However, whereas FeZnO maintains a stable conversion around 40% in the subsequent cycles, FeCuO continues to deactivate from 63% down to values close to 40%. This indicates that, in absolute terms, FeCuO suffers a greater overall loss of activity as compared to FeZnO.

These results clearly highlight the reduced stability of both oxides under repeated reuse, which may be associated with catalyst deactivation phenomena such as leaching of active metals or accumulation of organic residues on the surface. With regard to the selectivity towards benzaldehyde, FeCuO displays an increasing preference for benzaldehyde over benzoic acid with each cycle, reflecting a progressive loss of oxidative capacity. Specifically, the selectivity to benzaldehyde rises from 81.1% to 100% after three cycles, consistent with a gradual deactivation of the catalytic system.

Evidence of metal leaching is also suggested by the colour change of the reaction mixture observed during the recycling experiments (Figure 9a), shifting from a coppery tone to colourless (see Eppendorf pictures). This behaviour is in agreement with previously reported describing the tendency of Cu to leach during heterogeneous oxidation reactions, negatively affecting both activity and durability of the catalyst [53]. To further investigate this phenomenon, an additional test was carried out under the same conditions used in the reusability study (150 °C, 30 min, BA/TBHP 1:4), employing the reaction mixture recovered from FeCuO but without the heterogeneous solid confirming that leached copper was present in solution and was partially responsible of benzyl alcohol oxidation.

In parallel, catalyst deactivation due to incomplete removal of organic by-products over the catalyst surface during the cleaning procedure after three catalytic cycles was examined by FTIR-ATR (Figure 10). For this purpose, the materials were recovered by centrifugation (5 min at 1500 rpm), washed twice with the reaction solvent and twice with acetone, and then dried at 100 °C for 30 min.

In all spectra, bands at around 600 cm⁻¹ are observed, corresponding to metal–oxygen vibrations as previously discussed. In both FeCuO and FeZnO, the disappearance of the band at 1750 cm⁻¹, assigned to C=O stretching, is evident. This observation suggests that the surface structure of the catalysts undergoes modification upon reuse, which in turn impacts their catalytic performance. One possible consequence is the alteration or blocking of active sites, leading to a progressive loss of activity as the number of reuse cycles increases, mostly caused by humins on the catalytic surface due to high temperatures achieved during the process [57].

3. Materials and Methods

3.1. Materials

All reagents used for both the synthesis of the catalysts and the catalytic reactions were employed without further purification and were purchased from different suppliers depending on availability.

For the catalyst synthesis, the following chemicals were used iron (III) chloride hexahydrate (FeCl₃·6H₂O, Panreac, Barcelona, Spain), copper (II) chloride dihydrate (CuCl₂·2H₂O, Merck, Darmstadt, Germany), zinc chloride (ZnCl₂, purity 96%, Probus, Badalona, Spain), terephthalic acid (H₂BDC, C₆H₄ (CO₂H)₂, purity 98%, Sigma Aldrich, St. Louis, MO, USA), N,N-dimethylformamide (DMF, (CH₃)₂NCHO, purity 99.5%, Panreac), methanol (CH₃OH, Panreac), and hydrochloric acid solution 0.1 M (HCl, prepared from 35% stock, Panreac).

For the catalytic reactions, the following were used: tert-butyl hydroperoxide (TBHP, 70% in water, Sigma Aldrich) as oxidant, benzyl alcohol (C₆H₅CH₂OH, purity 99%, Sigma Aldrich) as substrate, and acetonitrile (CH₃CN, Panreac) as solvent.

3.2. Catalyst Synthesis: MOFs Synthesis

MOF materials were obtained via the typical solvothermal route. As a reference, the synthesis protocol carried out in our research group for Fe-MIL-101 was followed [15,58]. Specifically, 1.35 g (5 mmol) of FeCl₃·6H₂O and 1.678 g (10 mmol) of H₂BDC were dissolved in 40 mL of DMF and 0.5 mL of 0.1 M HCl, since an acidic medium promotes the formation of MOF nuclei. Subsequently, the mixture was stirred for 1 h at 300 rpm, then the homogeneous solution was transferred to a 125 mL Teflon-lined autoclave (BERGHOF, Eningen, Germany) and heated at 150 °C for 24 h. The resulting solids were recovered by filtration, washed twice with methanol, and dried overnight in an oven at 80 °C. In Figure 11 is depicted the scheme of MOFs synthesis (step 1 to 4).

An analogous procedure was applied for the synthesis of bi- and trimetallic MOFs, adjusting the amounts of metal precursors according to the desired composition. The synthesis was repeated in parallel as many times as necessary to obtain sufficient material for both characterisation and subsequent catalytic evaluation. The amount of metal precursors and organic linker used for the synthesis of MOFs and the corresponding metal oxides are summarised on

Table 4. Molar relation and precursors used for MOFs synthesis.

MOF	Metallic Precursor	Weight (g)	Mmol	Organic Linker	Weight (g)	Mmol
Fe-MIL-101	FeCl3	1.350	5	Terephthalic acid	1.678	10
FeCu-MIL-101	FeCl3 CuCl2	0.670 0.426	2.5
FeZn-MIL-101	FeCl3 ZnCl2	0.670 0.341	2.5
FeCuZn-MIL-101	FeCl3 CuCl2 ZnCl2	0.373 0.284 0.227	1.67

.

3.3. Metal Oxide Synthesis

Metal oxides were obtained from the previously synthesised mono-, bi-, and trimetallic MOFs. Once dried (step 4, Figure 1), the materials were subjected to thermal treatment in a muffle furnace under a controlled air/N₂ atmosphere. The process consisted of calcination at 600 °C for 2 h, with a heating ramp of 4 °C/min. This treatment allowed the decomposition of the organic ligand and the formation of the corresponding metal oxides.

The nomenclature used throughout this work to identify both the MOFs and their derived oxides is presented in

Table 5. Nomenclature of catalysts obtained: MOFs and oxides.

MOF	Code	Oxide	Code
Fe-MIL-101	Fe-MOF	Fe2O3@C	FeO
FeCu-MIL-101	FeCu-MOF	CuO-Fe2O3@C	FeCuO
FeZn-MIL-101	FeZn-MOF	ZnO-Fe2O3@C	FeZnO
FeCuZn-MIL-101	FeCuZn-MOF	CuO-ZnO-Fe2O3@C	FeCuZnO

, in order to facilitate the interpretation of the results.

3.4. Catalyst Characterization

The catalysts were characterised to determine their textural, structural, and physicochemical properties, which provide insight into their composition, internal organisation, and behaviour under different reaction conditions. For this purpose, several complementary analytical techniques were employed, including N₂ adsorption–desorption porosimetry, X-ray diffraction (XRD), attenuated total reflectance Fourier transform infrared spectroscopy (FTIR-ATR), transmission electron microscopy (TEM), and scanning transmission electron microscopy (STEM) coupled with an energy-dispersive X-ray spectroscopy (EDX) detector from the research support service SCAI (University of Córdoba).

3.4.1. N₂ Adsorption-Desorption Porosimetry

Textural properties of the porous solids were determined by nitrogen adsorption–desorption at −196 °C (liquid nitrogen). Measurements were carried out using a Micromeritics ASAP 2000 analyzer (Norcross, GA, USA). Prior to analysis, 0.1–0.2 g of each sample was degassed at 0.1 Pa for 24 h at 130 °C. Specific surface area was calculated from the linear region of the BET equation, assuming a nitrogen molecular cross-sectional area of 0.162 nm². Total pore volume and mesopore distribution were evaluated using the BJH method.

3.4.2. X-Ray Diffraction (XRD)

XRD analysis was used to investigate the crystallinity and structure of the solid materials. Diffraction patterns were recorded on a Bruker D8 DISCOVER diffractometer (Billerica, MA, USA) using Cu Kα radiation (λ = 1.5418 Å) over the 2θ range of 5–80° with a scan rate of 1°/min.

3.4.3. FTIR-ATR Spectroscopy

FTIR-ATR measurements were performed to obtain information on the surface chemical composition of the samples. Spectra were recorded on a PerkinElmer Spectrum 3 spectrophotometer (Boston, MA, USA) equipped with a PIKE Technologies MIRacle™ single reflection accessory. Measurements were carried out in air over the range 4000–600 cm⁻¹ with a resolution of 8 cm⁻¹, averaging 60 scans for each spectrum.

3.4.4. Transmission Electron Microscopy (HR-TEM and STEM)

TEM analysis was used to investigate the morphology of the synthesised materials at the atomic scale. High-resolution TEM images were obtained with a Talos F200i microscope (Waltham, MA, USA) available at the Central Research Support Service (SCAI) of the University of Córdoba.

STEM imaging was performed with the same instrument equipped with an energy-dispersive X-ray spectroscopy (EDX) detector, enabling elemental mapping and distribution analysis at the nanoscale.

3.5. Catalytic Tests

The oxidation of benzyl alcohol to benzaldehyde was carried out using a CEM Discover 2.0 microwave reactor(Charlotte, NC, USA) from our research group. Reactions were performed in 10 mL Pyrex glass vials. Each reaction contained 50 mg of catalyst, 0.1 M benzyl alcohol (0.021 mL), 0.4 M tert-butyl hydroperoxide (0.111 mL), and acetonitrile (ACN) as solvent until a final volume of 2 mL. Reaction conditions were varied depending on the specific objective, with microwave power fixed at 300 W, temperature in the range of 90–150 °C (2.5 min heating ramp), and reaction time between 5 and 45 min.

Product identification was carried out by gas chromatography–mass spectrometry (GC–MS) using an Agilent 7820A GC/5977B (Santa Clara, CA, USA) instrument equipped with a 30 m × 0.25 mm × 0.25 μm HP-5ms capillary column. Analysis conditions were as follows: injector temperature 250 °C, detector temperature 300 °C, oven temperature initially 80 °C and ramped at 10 °C/min to 230 °C, giving a total analysis time of 42 min.

Quantification of the identified compounds was performed using an Agilent 5890 Series II gas chromatograph equipped with a flame ionisation detector (FID) and a non-polar Supelco Equity™-1 silica column (60 m × 0.25 mm × 0.25 μm). The injector and detector were maintained at 250 °C. The oven programme started at 60 °C and increased at 5 °C/min to 180 °C, with a total analysis time of 26 min.

Conversion, selectivity, and yield were calculated according to the following expressions:

$$Conversion\left(\%\right)={\displaystyle \frac{n_0-n_f}{n_0}}·100$$

$$Selectivity\left(\%\right)={\displaystyle \frac{n_p}{n_t}}·100$$

$$Yield\left(\%\right)={\displaystyle \frac{n_p}{n_0}}·100$$

where n₀ is the initial mola amount of benzyl alcohol, n_f the molar amount resting of benzyl alcohol and n_p is the molar amount of the desired product.

Carbon balance was assessed by GC-FID quantification, obtaining a loss-value in all cases lower than 2% of carbon loss. The carbon balance was estimated from the GC–FID analysis of the liquid phase, considering the carbon content of benzyl alcohol, benzaldehyde and benzoic acid. Peak areas were used directly to determine the relative molar amounts, assuming a similar FID response per mole of carbon for these aromatic compounds.

4. Conclusions

In this work, the catalysts FeO, FeCuO, FeZnO, and FeCuZnO were successfully synthesised from their MOF precursors: MIL-101 (Fe), MIL-101 (FeCu), MIL-101 (FeZn), and MIL-101 (FeCuZn). The structure of these materials was confirmed by a convenient selection of characterisation techniques, and their performance was evaluated in the selective oxidation of benzyl alcohol to benzaldehyde. Among the different conditions tested, the best results were obtained at 150 °C and 30 min, with FeZnO achieving the highest overall yield (62.1%).

The superior performance of FeZnO compared to the other catalysts can be explained by its physicochemical properties. STEM images revealed that Zn is mainly located at the surface, while N₂ adsorption–desorption measurements showed that FeZnO has the highest surface area and pore volume among the oxides. Together, these features result in better accessibility to the active sites, enhancing catalytic activity. In contrast, the copper-containing catalysts (FeCuO and FeCuZnO) displayed lower structural stability, with an evident Cu leaching during the reaction.

Although Cu leaching could not be quantified due to temporary malfunction of the analytical equipment, its occurrence was inferred from the discoloration of the reaction medium and the progressive loss of activity over successive cycles, consistent with literature reports for Cu-based catalysts under oxidative conditions. Future studies will focus on quantifying this effect and improving Cu retention within the oxide matrix.

Reusability studies showed that both FeCuO and FeZnO gradually lose activity with increasing cycles. Although FeZnO deactivates more rapidly, it still maintains a higher overall yield as compared to FeCuO. The observed Cu leaching provides a plausible explanation for the faster deactivation of copper-based catalysts, consistent with previous literature.

Overall, these findings confirm that MOF-derived oxides are more active and selective than their respective parent MOFs, in addition to the improvement in the catalytic performance achieved by bi- and trimetallic systems. However, the catalysts stability remains a challenge, particularly in Cu-containing materials. Future work will focus on optimising the synthesis routes to stabilise copper within the oxide framework and minimise leaching, while also exploring new strategies to enhance catalyst durability and broaden their application to the selective oxidation of other biomass-derived alcohols. Additionally, the combination of MOF-derived oxides with microwave-assisted processes represents a promising and energy-efficient route for sustainable fine chemical synthesis. Nonetheless, further studies on scaling-up and long-term catalyst stability are needed to assess the system’s practical feasibility.