Copper Molybdate-Catalyzed Esterification of Levulinic Acid: A Heterogeneous Approach for Biofuel Synthesis

Abstract

The catalytic esterification of levulinic acid (LA) to methyl levulinate (ML) was investigated using copper molybdate (Cu₃(MoO₄)₂(OH)₂) as a heterogeneous catalyst. The catalyst, synthesized via chemical precipitation, exhibited a monoclinic structure with self-assembled nanoplates forming spherical mesostructures. Structural characterization confirmed its high crystallinity, while textural analysis revealed a BET surface area of 70.55 m² g⁻¹ with pore sizes in the nanometric range (1–6 nm). The catalytic performance was systematically evaluated under varying reaction conditions, including temperature, catalyst dosage, reaction time, methanol-to-LA molar ratio, alcohol type, and catalyst reusability. Optimal conversion of 99.3% was achieved at 100 °C, a 1:20 methanol-to-LA molar ratio, 5% catalyst loading, and a reaction time of 4 h. Comparative analysis with other heterogeneous catalysts demonstrated superior efficiency and stability of Cu₃(MoO₄)₂(OH)₂, with minimal activity loss over four reuse cycles (final conversion of 77.1%). Mechanistic insights suggest that its high activity is attributed to Lewis and Brønsted acid sites, facilitating efficient esterification. This study underscores the potential of copper molybdate as a sustainable and recyclable catalyst for biofuel additive synthesis, advancing green chemistry strategies for biomass valorization.

1. Introduction

The ecological impacts caused by the predatory use of natural reserves have become constant and reported by climate change that has affected the entire world, causing an increase in the volume of the oceans, droughts, storms, desertification, and mass extinction of species [1,2]. In this context, the future of the next generations and the existence of the species of animals, plants, and other organisms that make up our ecosystem are entirely compromised by the attitudes taken throughout this decade. Therefore, the change in human behavior, especially concerning the use of renewable energy matrices, reduction of consumerism and, consequently, the volume of solid waste generated, partial or total replacement of petroleum inputs, and maintenance of water resources, are some of the emerging initiatives [3,4].

According to the World Economic Forum (WEF), the world population has reached 8 billion people and is estimated to reach 10.4 billion people in the year 2086 [5]. These data reinforce the need for strategic strengthening according to the sustainable development goals of the United Nations (UN), especially concerning the quality of water resources, affordable and clean energy, healthy life, and maintenance of terrestrial ecosystems and biodiversity. Based on these premises, the search for renewable and biogenic sources of inputs for various purposes has been the subject of studies by numerous researchers worldwide, who aim to partially or entirely replace traditional inputs with biodegradable materials from residual matrices [6,7,8]. This approach allows, among other main advantages, the reduction in dependence on petroleum inputs, valorization of production and extractive chains and, consequently, preservation of forests and green areas, reduction in the volume of urban solid waste, and reduction in emissions of compounds harmful to the environment [9,10,11].

Vaithyanathan et al. [12] report the importance of valuing research related to the economic valorization associated with the transformation of residual biomass into products of industrial interest, also called commodity chemicals, which allows for economic gain from solid waste composed of biomass-derived materials, mainly from extractive activity, agriculture, and wood derivatives. In addition, the authors present a classification model for residual biomass according to the source of the raw material and the advantages and disadvantages of these inputs for energy generation. The first generation of biomass includes starch and vegetable oils and animal fats, i.e., mono-, di-, and triglycerides, which are used to obtain biofuels, resins, polymers, cosmetics, inputs for the processing industry, and agrochemicals. The second generation includes, among others, cellulosic lignin biomass and organic residues from the processing or post-consumption of food, agroforestry residues, and manure, which can be used in the production of biogas, biofertilizers, animal feed, and inputs to produce biopolymers, biochar, and chemical inputs for the transformation industry. The third generation concentrates biomass derived from algae, both micro-, and macroalgae, which has the advantage of a rich availability of polysaccharides, enzymes, and oils, which provide opportunities for the development of cosmetics, biopolymers, biofuels, and other materials under lower or no concentrations of pesticides.

The study of the transformation of residual biomass into molecules of industrial interest has become the object of study over the last decades, mainly driven by investments motivated by bioeconomy. In this way, it is possible to highlight the chemical transformation of vegetable fibers, rich in lignin, cellulose, and hemicellulose, into molecules called platform, as is the case of 5-hydroxymethylfurfural—HMF [13], formic acid [14], alcohols [15], sugars [16], sorbitol and mannitol [17], lactic acid [18], and levulinic acid [19].

Among the compounds presented in the previous paragraph, levulinic acid (C₅H₈O₃), also known as 4-oxopentanoic acid, in addition to other organic acids, has aroused the interest of the industrial due to its applications in the field of agrochemicals, biofuels, pharmaceuticals, cosmetics, and plasticizers and adhesives [20]. Several chemical routes are presented in the literature to obtain levulinic acid, among which it is possible to highlight the use of biorefinery, starting from sugarcane biomass [21], rice hull [21], straw [22] and corn cobs [23], potato peel [24], hazelnut shell [23], and coconut shell [25]. Besides that, levulinic acid is also used in different reactions to obtain different products, such as ketalization [26], esterification [27], amination, oxidation/reduction [28], and condensation [29], among others. In the vast majority, homogeneous or heterogeneous catalysts are used to increase the percentage of conversion, selectivity, and/or reduction of the reaction period required in the processing, commonly adopting hydrothermal or microwave-assisted hydrothermal methods [26].

The literature reports recent studies [13,21,22,30,31,32] in the field of biofuel applications that sought to obtain methyl esters from levulinic acid using homogeneous and heterogeneous catalysis. In this regard, Chaffey et al. [31] obtained biofuel esters and lactones from levulinic acid using ZSM-5 and Amberlyst^®-15 materials as heterogeneous catalysts (catalyst dosage = 40 mg, 1 mmol of DMPO in 2 mL of DMC as solvent at 75 °C for 1 h), resulting in a conversion percentage higher than 90%, even after the third cycle of reuse. On the other hand, Zhao et al. [30] used a sulfonic acid-functionalized lignin-montmorillonite complex as a catalyst in the esterification reaction of levulinic acid into n-butyl levulinate, which resulted in a conversion percentage close to 98% in hexane as a solvent, using catalyst dosage, temperature, and reaction time of 10%, 120 °C, and 4 h, respectively. Gallego-Villada et al. [33] also performed the esterification of n-butyl levulinate using n-butanol and Preyssles’s catalyst—H₁₄(NaP₅W₃₀O₁₁₀); in this study, the authors obtained a yield of 77% (mol) conversion and 100% (mol) selectivity over 3 h of reaction at a temperature of 160 °C. Therefore, it is noted that the search for catalysts that provide reactions for the esterification of levulinic acid has been a promising field of research aimed beyond obtaining biofuels and other types of molecules of industrial and commercial interest.

It is known that the class of molybdates (M²⁺MoO₄) and tungstates (M²⁺WO₄), where M²⁺ is a metal with a 2+ valence, are promising materials for different applications, including photocatalysis [34], antimicrobials [35], electrochemical [36], capacitors [37,38], diodes [39], and sensors [36], and as heterogeneous catalysts in esterification and transesterification reactions [40]. Among the transition metal molybdates (M²⁺ = Cu²⁺, Ni²⁺, Fe²⁺, Zn²⁺, and Mn²⁺), the copper molybdate polymorphs, which exhibit the formulas CuMoO₄, Cu₃(MoO₄)₂(OH)₂, Cu₃(MoO₄)(OH)₄, and Cu₃O(MoO₄)₂, have excellent physicochemical properties that make them promising in catalytic applications. In addition, six polymorphs are currently known for copper molybdate, exhibiting five crystal structures, where the alpha phase (α-CuMoO₄) is metastable, and the alpha-beta phase (β-CuMoO₄) is predominant at high temperatures. However, variations, called gamma phase (ϒ-CuMoO₄), Type II (CuMoO₄-II), and III (CuMoO₄-III), are also possible, above all, under specific pressure conditions and temperatures. The polymorphs Cu₃(MoO₄)₂(OH)₂ and Cu₃(MoO₄)(OH)₄ exhibit natural occurrence, these being popularly known as the Lindgrenite minerals [41,42] and Markascherite [43], respectively.

Lindgrenite exhibits a monoclinic structure with a space group P21/n and two formulas per unit cell, the result of the organization of cluster’s [MoO₄] tetrahedral symmetry, and the clusters [CuO₆]OH or [CuO₄(OH)₂], the octahedral symmetries of which suffer substantial bond distortions in octahedral clusters mainly due to the Jahn–Teller effect and the configuration of copper ions in the structure. The synthesis method and applications of synthetic lindgrenite, that is, its catalytic, photoluminescent, and antimicrobial properties, have been previously investigated [41,42] and were confirmed in all of the applications investigated with high performance and reproducibility of the results obtained.

Herein, we investigate the esterification of levulinic acid under different experimental conditions, utilizing copper molybdate nanocrystals in the form of synthetic lindgrenite as a heterogeneous catalyst. The nanostructured catalyst was synthesized via chemical precipitation at room temperature for 90 min, resulting in highly crystalline and well-defined morphologies. The catalytic performance was systematically evaluated by varying key reaction parameters, including temperature, catalyst loading, reaction time, and the molar ratio of reactants. Furthermore, the catalyst stability and reusability were evaluated over multiple cycles to further assess its durability and potential for sustainable industrial applications.

2. Results and Discussion

2.1. Structural and Morphological Characterization

Figure 1a,b presents the structural characterization by X-ray diffraction (XRD) and vibrational Raman spectroscopies performed with the copper molybdate sample synthesized by the chemical coprecipitation method. Indexing the crystallographic planes for the data collected in the interval 2θ between 10 and 80° (Figure 1a) reveals the characteristic profile of materials with a high degree of crystallinity and ordering [35,44,45]. In addition, the occurrence of a single phase is confirmed, named tricopper bis(molybdate) dihydroxide and chemical formula Cu₃(MoO₄)₂(OH)₂, following the literature consulted [41,42,43,46]. Therefore, the identified structure exhibits crystallographic plans characteristic of the monoclinic, special group structure P21/n; no diffraction peaks associated with the presence of synthesis precursors or secondary phases were evidenced, that is, XRD diffraction peaks associated with the copper oxide (CuO) or molybdenum oxide (MoO₃), which confirms the efficiency of the synthesis method adopted. The comparison of the crystallographic information for copper molybdate obtained in this study with the crystallographic standard contained in the inorganic crystal structure database (ICSD) no. 30946 confirms the obtaining of copper molybdate with network parameters of a = 5.39(4) Å, b = 14.02(3) Å, and c = 5.60(8) Å, and a unit cell volume (V) of 419.53(1) ų [47,48,49].

The formation of copper molybdate with a structure similar to the mineral lindgrenite resulted primarily from the ionic interaction between the ions Cu²⁺ and $\mathrm{M}\mathrm{o}{\mathrm{O}}_4^{2-}$, which had, as precursors, the salts Cu(NO₃)₂∙3H₂O and Na₂MoO₄∙2H₂O, respectively [41,46]. In aqueous media, copper nitrate solubilizes easily (137.8 g/100 H₂O, 0 °C), a behavior that was also observed for sodium molybdate dihydrate (84 g/100 g H₂O, 100 °C). The crystallization of the structure leads to the formation of tetrahedral clusters composed of molybdenum atoms coordinated with four oxygen atoms [MoO₄], while octahedral symmetry clusters [CuO₄(OH)₂] are composed of copper atoms coordinated with six oxygen atoms, which have the presence of hydrogen atoms for two of the shared oxygens, resulting in hydroxyl groups [36].

Corroborating the information obtained by X-ray diffraction, the Raman spectroscopy, as shown in Figure 1b, resulted in molecular vibration bands in the interval between 100 and 1000 cm⁻¹, which are in excellent agreement with the spectra reported in the RRUFF website for copper molybdate minerals with monoclinic structure, as found by John Dagenais in Chuquicamata mine, Calama, Antofagasta Province, Chile (R050213); Rock Currier, in San Samuel mine, Carrera Pinto, Chile (R060241); and Ron Gibbs, in Childs Aldwinkle mine, Copper Creek, Pinal County, Arizona, USA (R110061). From the spectrum presented in Figure 1b, it was possible to identify the presence of the vibration modes in the wave numbers 984, 931, 901, 885, 840, 796, 775, 497, 401, 361, 337, 313, 302, 289, 254, 217, 174, 129, and 105 cm⁻¹. These vibrational modes are associated with the vibrations of the units [MoO₄] and [CuO₄(OH)₂] [43]. In particular, the strong intensity bands of the vibration modes associated with the symmetrical and asymmetrical stretches of the Mo-O connections of the clusters [MoO₄], characteristic of transition metal molybdates [30,47,48,49,50,51,52,53], were identified in the values of 933 and 901, respectively.

In the study conducted by Wen et al. [48], polymorphs of copper molybdate were obtained using the hydrothermal method, starting from ammonium molybdate and copper nitrate in an aqueous medium. The authors confirm, from the Raman vibrational spectroscopy data, the presence of bands at the values 289, 304, 342, 363, 405, 797, 841, and 930 cm⁻¹; in addition, they report that the band located in 289 cm⁻¹ is associated with the stretching mode of Cu–O bonds along the crystal lattice. The bands at 304, 342, 363, and 405 cm⁻¹ correlate with the symmetrical and asymmetric stretches of the O–Mo–O bonds present in the clusters [MoO₄] of tetrahedral symmetry. The bands at 841 and 930 cm⁻¹ are due to the symmetrical and asymmetrical stretches of the Mo–O connections; finally, the band in 930 cm⁻¹ is due to the vibration of the Mo=O groups and is considered an indicator of the presence of oxygen vacancies and different valences for molybdenum.

Figure 2a,b shows the Fourier transform infrared (FTIR) spectra of copper molybdate synthesized by the chemical precipitation method, as well as the spectrum of copper molybdate before and after exposure to pyridine adsorption for determination of Brønsted and Lewis acid sites. Starting from the graphical analysis of the spectrum presented in Figure 2a, it is possible to corroborate the information given by Raman spectroscopy, where the molecular vibration bands observed in the range from 480 cm⁻¹ to 4000 cm⁻¹ are all assigned to the structure of the Cu₃(MoO₄)₂(OH)₂. Therefore, the vibrations associated with the bands at 866 and 918 cm⁻¹ are attributed to the asymmetric stretches, while the band at 972 cm⁻¹ is associated with symmetrical stretches of the Mo–O bonds of the tetrahedral clusters, respectively [54]. Vibrations identified in the range between 3250 and 3500 cm⁻¹ are due to stretches of the O–H bonds present in octahedral clusters [CuO₄(OH)₂] [55]. The asymmetric band of strong intensity between 774 and 866 cm⁻¹, is due to the vibrations of the crystalline lattice for unity Cu₃[MoO₄]₂O.

The spectrum in Figure 2b presents the graphic profile of the modes of molecular vibration of copper molybdate before and after exposure to pyridine molecules, which allows the identification of the Brønsted and Lewis sites. In this case, the interaction of pyridine molecules results in the appearance of bands at the interval between 1425 and 1650 cm⁻¹ [31]. Vibration bands near 1451 cm⁻¹ and 1543 cm⁻¹ are generally associated with the Lewis and Brønsted sites, respectively [56]. In this study, it is noted that bands emerged at 1427 and 1590 cm⁻¹, which confirms the presence of Lewis and Brønsted sites. The first case is confirmed by the presence of the band in 930 cm⁻¹ in the Raman spectrum, which confirms the presence of molybdenum atoms with the configurations Mo⁶⁺, Mo⁵⁺, and Mo⁴⁺. On the other hand, the Brønsted sites are associated with the OH groups present in the clusters [CuO₄(OH)₂] of octahedral symmetry.

The thermal stability of copper molybdate was investigated by adopting the thermogravimetric analysis, in this case, by means of the thermogravimetry (TG), derived thermogravimetry (dTG), and differential scanning calorimetry (DSC) curves. In addition, the textural properties were studied by nitrogen adsorption/desorption (N₂), adopting the Brunauer–Emmett–Teller (BET) method. Figure 3a–d shows the characterizations obtained by TG/dTG (Figure 3a), DSC (Figure 3b), adsorption/desorption hysteresis of N₂ (Figure 3c), and pore distribution, as shown in Figure 3d.

The analysis of the graphic data presented in Figure 4a reveals three main mass loss events for copper molybdate in the ambient temperature range up to 900 °C. In this case, the first event, observed in the DTG curve at 295 °C, which implied a weight variation of 5.76%, was also evidenced in the DSC curve at 303 °C (Figure 4b); this is due to the release of water molecules, a process that occurs under energy absorption, that is, endothermically, and is characterized by the reactivity between the hydroxyl groups present in the clusters [CuO₄(OH)₂]. Consequently, it results in a change in the monoclinic structure of the Cu₃(MoO₄)₂(OH)₂ in the orthorhombic structure of the Cu₃Mo₂O₉, which has a space group Pna21 [57]. The second event was observed at a temperature of 502 °C, as can be observed in the TG/DTG curve, which implied the event of low intensity in the DSC curve, more precisely, at 520 °C, and is due to the process of recrystallization of the structure of the Cu₃Mo₂O₉, a characteristically exothermic process. Finally, the third event, associated with the decomposition of the Cu₃Mo₂O₉ in copper oxide and molybdenum trioxide (MoO₃), occurred at a temperature of 820 °C (TG/DTG), resulting in an endothermic event, as shown in Figure 4b, at a temperature of 822 °C, which varied by 2.45% of the sample weight. The two mass losses that occurred at temperatures of 520 °C and 822 °C may be related to the formation of oxygen atom vacancies, which, during the heating process and reorganization of the structure, can carry O₂ molecules, implying the emergence of oxides with the metals molybdenum and copper, with different valences [56].

The textural properties of synthesized copper molybdate were studied by nitrogen adsorption/desorption (N₂), as shown in Figure 3c,d. Therefore, according to the IUPAC international classification, the hysteresis exhibited for the relative pressure (P/P₀) range from 0 to 1 is type III, i.e., materials absent from micropores, which corroborates the graph shown in Figure 3d. Therefore, based on these results is confirmed the occurrence of pores with radii in the predominantly nanometric range (1 and 6 nm) and specific surface area determined of 70.55(2) m² g⁻¹.

The morphological and semiquantitative analysis of the synthesized copper molybdate was investigated by scanning electron microscopy (SEM) and X-ray dispersive spectroscopy (EDX), as can be seen in parts of Figure 4a–c. Therefore, the morphology obtained for the synthesized material consists of the arrangement of nanocrystals in the shape of sheets, which result in spherical-shaped mesostructures with a diameter close to 9 μm. This arrangement is made possible by the self-assembly of the nanoplates of Cu₃(MoO₄)₂(OH)₂, which, through surface interactions, mainly involving hydroxide groups and differences in charges, promote the organization, as can be seen in Figure 4a,b. Moreover, the analysis by EDX revealed the presence of all the energy peaks associated with the characteristic elements of copper molybdate, namely, molybdenum, copper, and oxygen. For the peak of EDX for carbon, the carbon tape used to attach the samples to the sample holders (stubs) is due.

In the study conducted by Wen et al. [48], irregular agglomerates of nanoplates of Cu₃(MoO₄)₂(OH)₂ were obtained by the hydrothermal method, using different synthesis conditions. While Wu et al. [58] report the achievement of spherical structures of Cu₃(MoO₄)₂(OH)₂ with dimensions between 300 and 500 nm, in this case, adopting the method of chemical precipitation with the addition of hydrazine, at a temperature of 70 °C. The study carried out by Carvalho et al. [41] also reports the obtaining of spherical-shaped mesostructures, composed of nanoplates that underwent the self-assembly process, according to the authors, due to the surface interaction in the search to reduce the surface energy difference.

2.2. Catalytic Conversion of Levulinic Acid into Methyl Levulinate

2.2.1. Effect of Temperature

Increasing the temperature increases the reaction rate due to the higher kinetic energy of the molecules and promotes a reduction in solution viscosity. This improvement facilitates better interaction between the reactants and the catalyst [59]. The conversion of LA increases with temperature between 80 and 120 °C, reaching up to 99.33% (acid–base titration) and 96% (NMR) (Figure 5). This behavior is expected, as esterification is an endothermic reaction, which shifts the equilibrium toward the products at higher temperatures [60]. Similar optimal temperatures were reported in studies applying heterogeneous catalysts: carrageenan/DFNS composite (150 °C, 61% conversion, [61]); defective MOF-808 (74 °C, 97.5%; [62]); MOF/TiO₂ composites (100%, 120 °C [63]); and phosphotungstic acid supported on residual sugarcane bagasse (90 °C, 95%; [60]). Therefore, considering green synthesis and low energy consumption, 100 °C is deemed ideal [62].

2.2.2. Catalyst Dosage

The efficiency of the reaction is directly related to the availability of active sites in the reaction medium that promote product formation. In this study, the catalyst dosage varied from 1% to 7% wt. relative to the mass of levulinic acid. Figure 6 shows that, at 1% (wt.), the conversion was relatively low, similar to the reaction without a catalyst. This corroborates the idea that, at low dosages, the number of catalytic sites available on the lindgrinite cannot efficiently adsorb and activate all the reactant molecules. As the dosage increases, more active sites become available, facilitating adsorption and the reaction between levulinic acid and methanol [64]. At dosages of 5% wt. and above, the catalyst quantity exceeds the demand of the available reactants, leading to the saturation of the catalytic sites. This explains the stabilization of conversion from this dosage onwards. However, excessive catalyst dosages may reduce conversion due to increased solution viscosity, which hinders the esterification reaction. Excess catalyst impairs mass transfer, and the increased mixture viscosity may compromise catalytic activity [65]. Higher catalyst dosages can also generate more solid waste or require more energy for recycling and recovery. Therefore, 5% was considered the optimal catalyst dosage. Similar studies reported higher catalyst amounts, such as commercial TiO₂ nanoparticles (8.6% wt. of catalyst, 77.6% wt. conversion [66]); copper molybdate nanoplates (5% wt., 98.38% [42]), sulfonated corn cobs (10% wt. of catalyst, 83.15% [67]); and carbon cryogel (25% wt. of catalyst, 87.2% [68]).

2.2.3. Effect of Time

By varying the reaction time, it is possible to see a gradual increase from 1 to 3 h (Figure 7). At 4 h, maximum conversions were observed. By titration, a slight decrease was observed in the conversion measured by titration, but it remained the same when measured by ¹H NMR. This result shows that titration, despite being a cheaper technique, is more susceptible to systematic and operator-dependent errors than NMR, requiring careful alcohol evaporation from the medium to reduce errors. However, good reproducibility was observed in both techniques. Therefore, 4 h was considered the optimal reaction time. Higher optimal reaction times were observed in similar studies using heterogeneous catalysts: metal oxide-supported catalysts (TiO₂, SiO₂, and Al₂O₃) (6 h, 96% conversion; [69]); anchored silicotungstic acid (90 °C, 4 h; [70]); sulfonated corn cobs (9 h, 83% [67]).

2.2.4. Effect of Molar Ratio of Levulinic Acid to Methanol

The molar ratio between methanol and levulinic acid directly affects the equilibrium shift in the esterification reaction and the final yield. This behavior is evident in Figure 8, where, at lower initial molar ratios (5:1), the methyl levulinate (ML) yield is relatively low, indicating that the limited amount of methanol is insufficient to shift the equilibrium toward the product significantly [32]. As the molar ratio increases to 10:1 and 20:1, the conversion grows substantially, reaching 99% (NMR) at 20:1, which is considered the optimal point of the reaction. This observation aligns with Le Chatelier’s principle, as esterification is a reversible reaction, and increasing the concentration of one reactant (methanol) shifts the equilibrium toward forming the ester (ML). However, at a molar ratio of 30:1, the conversion decreases, showing that excessive methanol does not further promote product formation. This is likely due to the reaching of a saturation zone where the chemical equilibrium is already maximally shifted toward the product side [67]. In a related study, Alsalim et al. [61] used a carrageenan/DFNS composite catalyst under optimal reaction conditions (reaction time of 6 h, 10 mg catalyst, ethanol–LA molar ratio of 12:1, and reaction temperature of 150 °C) to achieve a 61% yield of ethyl levulinate.

2.2.5. Effects of Different Alcohol

The esterification reaction depends on the interactions between levulinic acid and alcohol at the acidic sites of the catalyst. More polar alcohols have a higher affinity for catalytic sites, facilitating the reaction. However, as the alcohol carbon chain increases, its polarity decreases, reducing the efficiency of the interaction. Figure 9 demonstrates this behavior, showing methanol achieving the highest conversion, likely due to its smaller molecular size and high polarity. In contrast, n-butanol exhibited lower yield, as its larger molecular volume and reduced polarity hinder its interaction with the catalyst [71]. Although ethanol has a shorter chain and is more polar than n-butanol, its conversion rate was lower. This peculiarity may be explained by its higher affinity for water, a byproduct of the reaction, leading to intermolecular associations that inhibit ethanol’s interaction with the catalytic sites [39]; in comparison, n-butanol, being less polar, is less affected by competition with water, allowing it to interact more effectively with the catalyst. Isopropanol, which showed the lowest conversion rate, and isobutanol are branched-chain alcohols that experience more significant steric hindrance, making it difficult to access the catalytic centers of lindgrinite. This reduces the overall product formation rate. Additionally, the molecular size of the alcohol affects the energy required to break and form bonds during the reaction, further lowering the yield [30]. Similar results were reported by Maggi et al. [72], who used sulfonic acids supported on silica for the esterification of levulinic acid with different alcohols, and Lobo et al. [59], who studied the esterification of LA and oleic acid using sulfonated passion fruit seed catalysts with various alcohols.

2.2.6. Recycling

As shown in Figure 10, the catalyst exhibits high initial efficiency in converting levulinic acid but undergoes progressive deactivation over successive recycling cycles. Nevertheless, it achieves more than 70% conversion after the fourth cycle, indicating its stability and viability for repeated use. Because of this result, we tested the possible leaching of the active phase in the reaction medium in optimized conditions. However, it did not happen in the methanol, confirming a total heterogeneous catalyzed reaction. Therefore, this decline in activity may be attributed to the formation of byproducts during the reaction, such as water or organic compounds, which may adsorb onto the catalyst’s active sites, blocking them and reducing their efficiency [62]. Additional factors contributing to decreased activity could include impurities accumulating in the catalyst’s pores, hindering reactant diffusion, or partial degradation of the catalyst structure due to repeated exposure to reaction conditions, such as high temperatures and the presence of reactants [69]. Similar behavior was observed in other studies. For example, Liu et al. [32] used a sulfonated solid carbon acid catalyst. They reported an 8% decrease in conversion, from 84.2 to 76.8%, after the first cycle under reaction conditions of a 5:1 EtOH/LA molar ratio, 0.3 g of catalyst, 120 °C, and 9 h. Their study attributed this decline to the possible instability of acidic sites that remained on the catalyst surface after being repeatedly washed with deionized water. Most of these unstable acidic sites were lost during the first cycle of the reaction.

3. Materials and Methods

3.1. Materials

Sodium molybdate dihydrate—Na₂MoO₄∙2H₂O (Sigma Aldrich, St. Louis, MO, USA, purity > 99.9%), copper nitrate trihydrate—Cu(NO₃)₂.3H₂O (Sigma Aldrich, purity = 98–104%) and sodium hydroxide—NaOH (Merk, St. Louis, MO, USA, purity > 98%), isopropyl alcohol—C₃H₈O (Synth, Rio de Janeiro, BR, purity = 95%), acetone—C₃H₆O (Synth, Rio de Janeiro, BR, purity > 99.5%), n-buthyl alcohol—CH₃(CH₂)₃OH (ACS reagent, St. Louis, MO, USA, purity ≥ 99.4%), 4-oxopentanoic acid—CH₃COCH₂CH₂COOH (Sigma Aldrich, St. Louis, MO, USA, purity = 98%), methanol—CH₃OH (Sigma Aldrich, St. Louis, MO, USA, purity ≥ 99.9%), pyridine—C₅H₅N (Sigma Aldrich, St. Louis, MO, USA, purity ≥ 99.4%), chloroform-d—CDCl₃ (Sigma Aldrich, St. Louis, MO, USA, purity = 99.8 atom % D, containing 1% (v/v) of TMS), ethanol—CH₃CH₂OH (Sigma Aldrich, St. Louis, MO, USA, purity ≥ 99.5%) were used without any previous treatment.

3.2. Methods

The synthesis of copper molybdate nanocrystals was conducted using the co-precipitation method as described by Carvalho et al. [41]. Figure 11 presents a detailed schematic diagram that clearly summarizes each step of the synthesis process. Initially, two solutions were prepared, named solution A and solution B, where solution A consisted of 50 mL of 4 × 10⁻³ mol of Na₂MoO₄∙2H₂O, while solution B consisted of 50 mL of 4 × 10⁻³ mol of Cu(NO₃)₂∙3H₂O. Solution A was subjected to constant magnetic agitation at 600 rpm. In contrast, solution B was added drop-by-drop, with a noticeable change in the suspension color obtained, intensifying the green color of solution B until its total transfer. The suspension remained for 90 min under constant magnetic agitation, followed by collecting the precipitate obtained by centrifugation, adopting 5 min cycles with a speed of 4000 rpm. At the same time, the spectator ions were efficiently removed by adopting successive washes with distilled water and, in the last cycle, with isopropyl alcohol. The collected material was dried in an oven for 24 h at 85 °C, followed by maceration, storage, and subsequent characterizations.

3.3. Characterization

The obtained materials were analyzed by X-ray diffraction (XRD) using a Shimadzu XRD-7000 diffractometer (Shimadzu Corporation, Kyoto, Japan) with a Cu/kα X-ray tube, λCu = 1.5406 Å, in the 2θ range between 10° and 100°, at a speed of 0.02 °min⁻¹, a voltage of 60 kV, and a current of 50 mA. The detailed crystallographic analysis of all samples was performed using the Rietveld refinement method. Therefore, in this study, the unit cell lattice parameters, atomic position, scale factor, and background were refined, in this case, adopting the pseudo-Voigt profile function in adjusting the intensity and profile of the experimental diffraction peaks.

Sample morphology and semiquantitative analysis were performed using scanning electron microscopy (SEM) and energy-dispersive X-ray spectroscopy (EDS) with a Tescan Vega3 microscope (Tescan Orsay Holding, Brno, Czech Republic), operating under high vacuum conditions. In this case, images were captured using a secondary and backscattered electron detector at different magnifications. Before the acquisition stage of the images, the samples were suspended in acetone, adding about 10 mg in 2 mL of acetone, followed by dispersing them in an ultrasonic washer of the brand Schuster, model L100, with a frequency of 42 MHz, followed by the collection of 100 µL of suspension and the pouring of it over the aluminum stub, initially containing the carbon tape, and then drying in an oven at a temperature of 35 °C for 1 h.

The vibrational Raman spectrum of the sample was collected using a Bruker confocal Raman spectrometer (SENTERRA model, Bruker Corporation, Billerica, MA, USA) in the spectral range of 50–1100 cm⁻¹. The sample was exciting using a green laser (λ = 532 nm) with output power, resolution, and integration time of 0.5 mW, 4 cm⁻¹, and 10s⁻¹, respectively.

The textural analysis by N₂ adsorption/desorption of the sample was conducted using a Quantachrome Autosorb iQ instrument (Boynton Beach, FL, USA) at −196 °C, within a relative pressure range from 0.005 to 1.0. This study used a 300 mg sample, which was previously degassed at 140 °C for 12 h under vacuum conditions and was subjected to the analysis. Based on this analysis, the surface area was estimated using the BET method (Brunauer, Emmett, Teller). The experimental data and data processing were acquired using ASiQwin software, version 5.21.

Vibrational analysis was performed using Fourier Transform Infrared (FTIR) spectroscopy with a Bruker Vertex 70 V spectrometer (Bruker Corporation, Billerica, MA, USA), employing the Attenuated Total Reflectance (ATR) technique. Thus, the IR data were collected using 32 scans with 4 cm⁻¹ of resolution in the range of 4000–400 cm⁻¹.

The determination of Lewis and Brønsted sites of the copper molybdate sample was achieved using the vibrational infrared spectrum of copper molybdate which was recorded using Bruker Fourier-transform infrared equipment. For this purpose, a portion of each dried sample was saturated with pyridine vapor, and the samples were exposed to pyridine in a controlled atmosphere for 24 h. Then, the wavenumber interval from 480 to 4000 cm⁻¹ was selected and the vibrational spectrum was recorded at a resolution of 4 cm⁻¹ under 32 scans.

3.4. Catalytic Esterification of Levulinic Acid

For the synthesis, 1 g of LA (8.61 mmol) was used in a steel-clad PTFE reactor with a volume of 15 mL, to which the solid catalyst and methanol were added in different proportions. The reactor was sealed and heated to the desired temperature with constant agitation. The reaction parameters were adjusted according to the experimental conditions, varying the temperature (80, 100, 120, and 140 °C), the reaction time (1, 2, 3, 4, and 5 h), the catalyst dosage (1, 3, 5, and 7% wt. related to the LA mass), the methanol/acid molar ratio (5, 10, 20, and 30 mol/mol), and the variation of alcohol (methanol, ethanol, isopropanol, isobutanol, and n-butanol). At the end of each reaction, an ice bath was used to rapidly cool the system, the products were centrifuged, and the excess alcohol was rotavaporated from the reaction medium using a rota evaporator (Buchi) with reduced pressure. Two different techniques determined the conversion of the product for comparison: (1) acid–base titration: the reaction product was dissolved in a mixture of solvents (ethanol–ether 1:1) and subsequently titrated with a solution of NaOH 0.1 molL⁻¹ and 1% (w/v) alcoholic solution of phenolphthalein as indicator [59]; (2) ¹H NMR spectrometry: a Bruker Avance Spectrometer III 300 MHz/54 mm was used to measure the conversion, and the product has been solubilized in CDCl₃. Based on Equation (1), we obtained the following conversion:

$$\mathrm{C}\mathrm{o}\mathrm{n}\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{s}\mathrm{i}\mathrm{o}\mathrm{n}\left(\mathrm{\%}\right)={\displaystyle \frac{{\mathrm{A}}_{{\mathrm{O}\mathrm{C}\mathrm{H}}_3}}{{\mathrm{A}}_{\mathrm{C}\mathrm{H}3}}}\times 100$$

(1)

where AOCH₃ is the integrated area referring to the hydrogens present in the methoxy group of the esters formed (3.67 ppm), and ACH₃ refers to the methyl group hydrogens (2.19 ppm) present in both the reagent and the product (Figure 12).

3.5. Catalyst Recycling and Leaching Test

Catalyst recycling was performed under the optimized reaction conditions (100 °C, 5% wt. of catalyst, 3 h of reaction). After each cycle, the solid catalyst was recovered by centrifugation, thoroughly washed with methanol under vigorous manual stirring to remove adsorbed reactants and products, and subsequently dried at 100 °C overnight in a forced-air oven prior to reuse in the following catalytic run.

In order to evaluate the potential leaching of the active phase into the reaction medium, a leaching test was performed, following the methodology proposed by Mendonça et al. [65] with slight modifications. In this procedure, the catalyst (5% wt) was suspended in methanol under the same optimized reaction conditions (100 °C, 3 h) in the absence of levulinic acid. After this period, the solid was separated by centrifugation, and the supernatant (methanol phase) was transferred to a clean reactor, to which levulinic acid was added at a methanol–acid molar ratio of 20:1. The mixture was then subjected again to the optimized reaction conditions (100 °C, 3 h), and the conversion was analyzed by the same method. The resulting conversion was negligible, indicating that no significant catalytic activity was present in the liquid phase, thereby confirming the heterogeneous nature and structural stability of the catalyst.

4. Conclusions

In summary, nanostructures composed of copper molybdate with a structure similar to the mineral lindgrenite were efficiently synthesized by the co-precipitation method at room temperature under constant magnetic stirring, starting from the reagents copper nitrate trihydrate and sodium molybdate dihydrate. Through the characterization techniques, it was possible to verify the formation of the monoclinic structure with the chemical formula Cu₃(MoO₄)₂(OH)₂, with morphology in the form of self-assembled nanoplates in spherical morphological mesostructures. The nanoplates showed adsorption/desorption hysteresis of type III N₂ isotherms, allowing the determination of the specific surface area by the BET method as 70.55 m² g⁻¹, and a pore distribution concentrated in the range from 1 to 6 nm. In addition, the thermal stability extended up to a temperature close to 300 °C, above which the material converted to the orthorhombic structure with the chemical formula Cu₃Mo₃O₉.

In the catalytic assays for the conversion of levulinic acid to methyl levulinate, a systematic study was conducted to evaluate the effects of temperature, catalyst dosage, time, methanol–alcohol ratio, and type of alcohol. Catalyst recycling was also investigated. The best experimental conditions were reached at a temperature of 100 °C, 5% catalyst dosage, 3 h of reaction time, and a methanol–acid molar ratio of 20:1, with methanol as the alcohol that led to the highest conversion (>97%). Meanwhile, catalyst reuse showed a gradual decrease in conversion, reaching approximately 77% after the fourth consecutive cycle of use, probably due to the adsorption of organic compounds on the catalyst surface. Leaching tests showed that the reaction occurs through a heterogeneous pathway and that the active phase does not leach into methanol after 4 h of reaction. Therefore, these results demonstrate the promising potential of copper molybdate nanostructures as efficient and recyclable heterogeneous catalysts for bio-based ester synthesis.