Introduction

Nowadays, the use of biomass-derived resources as renewable raw materials to produce chemicals and fuels has exponentially grown due to their potential ability to gradually diminish the dependence on fossil products [1, 2]. Platform molecules such as 5-hydroxymethylfurfural, furfural, and levulinic acid are examples of biorefinery compounds classified by the US Energy Department as “top 10” [3]. They are present in carbohydrates or lignocellulosic biomass and can be converted into a series of renewable fuels or chemicals [4,5,6].

Alkyl levulinates have physicochemical properties that allow their use as bioadditives of diesel fuel and gasoline or being added to the liquid fuel in the range of 0.5 to 20 vol% [7, 8]. These esters become even more attractive when the precursor alcohol has a renewable origin likewise ethyl alcohol [9].

In the conventional esterification routes, alkyl alcohol reacts with a carboxylic acid in the presence of a Lewis or Brønsted acid in a homogeneous phase [10,11,12,13,14]. However, liquid catalysts are hard to be separated from the products, are corrosive, and generally are not easily recyclable [15]. To avoid this drawback, acid catalysts can be supported on solid matrixes such as zeolites, anchored on ionic liquids, or as sulfonic resins [16,17,18,19,20].

Keggin heteropolyacids (HPAs) have been extensively used as catalysts for esterification reactions due to their advantages such as water tolerance, high acidity, and the possibility to be used as soluble or solid-supported catalysts [21,22,23,24,25,26,27,28]. Keggin heteropolyanions are clusters of metal–oxygen, where one XO4 central tetrahedral unit (i.e., X = P5+ or Si4+) is surrounded by W6+ or Mo6+ cations in an octahedral environment [29, 30].

The protons of Keggin HPAs are mobile and easily ionized in a polar medium, allowing a facile replacement by metal cations, generating active catalysts in different reactions in the homogenous phase [31,32,33,34]. However, Keggin HPAs can be also converted into heterogeneous catalysts, when the cations that replaced their protons have large ionic radium [35,36,37,38,39]. This partial replacement allows that the salts have still an acidity strength remaining, enough to catalyze esterification reactions under heterogeneous conditions, without compromising their catalytic activity. This approach has led the cesium HPAs to be used in different reactions such as oxidation, esterification, or etherification [40,41,42,43,44,45,46].

In this work, the focus was to assess the activity of cesium partially exchanged heteropolyacid salts in esterification reactions of levulinic acid, a biomass-derived platform molecule with alkyl alcohols. Effects of main reaction variables such as time, temperature, catalyst load, and type of alcohol were investigated. Among the Keggin HPAs evaluated, Cs2.5H0.5PW12O40 was the most active and selective toward alkyl levulinates. The reusability of the catalyst was successfully demonstrated.

Experimental section

Chemicals

All the chemicals and solvents were acquired from commercial sources and used without previous treatment. Levulinic acid and all the alkyl alcohols (i.e., methyl, ethyl, propyl, butyl, isopropyl alcohols) were Sigma-Aldrich (> 99 wt%). Phosphomolybdic, phosphotungstic and silicotungstic acids, and cesium carbonate were also Sigma-Aldrich (99 wt%).

Synthesis and characterization of the cesium HPA salts catalysts

The cesium salt catalysts were synthesized in agreement with the literature [40, 47]. Typically, an amount stoichiometric of Cs2CO3 aqueous solution was added dropwise to a solution of the Keggin HPA, aiming to achieve the molar ratio desired. The addition of Cs2CO3(aq) precipitated the Cs heteropoly salts as a white solid when phosphotungstic or silicotungstic acids were the precursors. When the phosphomolybdic acid was used, the cesium heteropoly salt was a green solid. HPA. After the water evaporation at 373 K, the Cs HPA salt was dried for 6 h in an oven. The characterization of the Cesium HPA salts used in this work was previously described in two previous works recently published [45, 48].

Identification of main reaction products

The reaction products were identified in a Shimadzu GC-2010 gas chromatographer coupled with an MS-QP 2010 mass spectrometer (i.e., electronic impact 70 eV, scanning range of m/z 50–450). Additionally, the products were co-injected with authentic samples synthesized as described in previous work [49, 50].

Catalytic tests

Tests catalytic were carried out in a sealed glass tube (25 mL). Typically, levulinic acid (2.0 mmol) was dissolved in an alkyl alcohol solution (8 mL) under magnetic stirring, and heated to 393 K in an oil bath. The addition of the acid catalyst (1.2 mol%) started the reaction.

The reaction progress was followed for 6 h, analyzing aliquots in GC equipment (Shimadzu 2010, FID), fitted with a Rtx®-Wax, capillary column (30 m length, 0.25 mm i.d., 0.25 mm film thickness) the aliquots periodically collected. The temperature program of GC analyses was as follows: 80 °C (3 min), heating rate (10 °C/min) until 240 °C. Injector and detector temperatures were 250 °C and 280 °C, respectively.

Results and discussion

Catalytic tests

Initially, the effect of the Keggin anion on the activity of Cesium salt catalyst was evaluated using ethyl alcohol as model alcohol following conditions described in the literature [49, 50]. In according with the literature, Cesium salts containing 0.5 mol of H+ ions/mol of catalyst have been the most efficient catalysts in different reactions [45, 48]. Therefore, all the Cesium HPA salts evaluated were tested with this load of H+ ions. Runs in the presence of the pristine heteropolyacid catalyst were also performed aiming for a comparison (Fig. 1).

Fig. 1
figure 1

Effect of Keggin anion on conversion and selectivity of esterification reactions of levulinic acid with ethyl alcohol. Reaction conditions: levulinic acid (2 mmol), ethyl alcohol (8 mL), dodecane (internal standard), catalyst (0.6 mol% of H+ ions), temperature (393 K), time (6 h)

It is important to highlight that all the catalysts were used at the same load of H+ ions. Another important point is that all the Keggin heteropolyacids are soluble in the reaction, whereas the insoluble Cesium salts. The soluble HPAs-catalyzed reactions achieved a minimum conversion of 97%. Although the Cesium silico-or phosphotungstic salts have a little bite less efficient than their precursor acids, this is still a satisfactory result since they are insoluble. The Cs2.5H0.5PMo12O40 salt achieved the lowest conversion (80%). Although the Cs3.5H0.5SiW12O40 salt has achieved a high conversion, it has a higher cesium content per mol of catalyst, therefore, the Cs2.5H0.5PW12O40 salt was selected to evaluate the other effects of reaction variables. The best performance of Cs2.5H0.5PW12O40 compared to their partner salts was assigned to the combination of its high surface area with strong strength of acidity, having been also observed in etherification reactions [45].

Fig. 1 shows that the selectivity of products formed in the esterification of levulinic acid with ethyl alcohol was less impacted by the nature of the catalyst. Ethyl levulinate (EL) was always the main product, while α-angelica lactone (αAL) and pseudo ethyl levulinate (PEL) were the secondary products (Scheme 1).

Scheme 1
scheme 1

Main products of levulinic acid esterification reaction with ethyl alcohol in the presence HPAs or their Cesium exchanged partially salts

Cesium phosphotungstate salts containing a variable load of H+ ions were evaluated and the main results are shown in Table 1.

Table 1 Conversion and selectivity of esterification reactions of levulinic acid in the presence of phosphotungstic acid and their cesium salts

In the range of concentration evaluated, the catalytic behavior of cesium exchanged partially salts were very close either in terms of conversion or selectivity toward ethyl levulinate. Although completely substituted with cesium, Cs3PW12O40 achieved still significant conversion and selectivity (Fig. 2). It suggests that this salt presents a strength of Bronsted acidity. Matachowski et al. assessed the acidity properties of a series of cesium phosphotungstate salts and assigned this remaining acidity to the hydrolysis of phosphotungstate by the hydration water molecule, which gives H+ ions [51]. This hypothesis was previously presented by Korosteleva et al. [52]. This can be the reason for the very similar activity of these salts (Table 1).

Fig. 2
figure 2

Kinetic monitoring of Cs2.5H0.5PW12O40-catalyzed levulinic acid esterification with ethyl alcohol. Reaction conditions: levulinic acid (2 mmol), ethyl alcohol (8 mL), dodecane (internal standard), catalyst (1.2 mol%), temperature (393 K)

The time reaction is a key aspect of this reaction; both conversion and EL selectivity were growing along the reaction, achieving the highest value at 6 h of reaction. The pseudo ethyl levulinate (PEL) was significantly formed within the first reaction interval (1.5 h). This product is obtained from α-ALac, which is formed from the cyclization followed by water molecule realising of levulinic acid (Scheme 1). Both five-membered ring products can be intermediate to give LE.

To evaluate the effect of catalyst load, the reactions were carried out with a variable amount of Cs2.5H0.5PW12O40 for 6 h. Fig. 3 shows the main results. After 6 h of reaction in the absence of catalyst, PEL was the major product. Although omitted herein, we have followed the selectivity of this reaction and verified that when LA is reacted with an excess of ethyl alcohol at 393 K, this is the major product since the reaction's beginning, which was gradually converted to LE with the reaction progress. However, as can be seen in Fig. 3, after 6 h PEL is still the main product.

Fig. 3
figure 3

Impact of catalyst load on the levulinic acid esterification with ethyl alcohol in the presence of the Cs2.5H0.5PW12O40 salt. Reaction conditions: levulinic acid (2 mmol), ethyl alcohol (8 mL), dodecane (internal standard), catalyst (1.2 mol%), temperature (393 K)

Besides increasing the conversion, the addition of the catalyst remarkably changes the reaction selectivity. Even at a minimum amount, the presence of the Cs2.5H0.5PW12O40 salt shifts the selectivity toward EL, evidence that it is essential to obtain the ester. A catalyst load higher than 0.6 mol% of H+ ions did not increase conversion or selectivity.

The impact of temperature was evaluated using the most favorable reaction conditions (0.6 mol% of H+ ions, 6 h reaction, Fig. 4). Conversion and selectivity were favoured by the increase in the reaction temperature. At higher temperatures, a greater number of reactant molecules can achieve the energy enough to overcome the activation energy of this reaction, increasing consequently its conversion. Moreover, since esterification reactions involve water molecules elimination this is an expected effect. The character endothermic of this reaction was previously demonstrated [45].

Fig. 4
figure 4

Impact of catalyst load on the levulinic acid esterification with ethyl alcohol in the presence of the Cs2.5H0.5PW12O40 salt

The activity of Cs2.5H0.5PW12O40 salt was also evaluated in the esterification of LA with other alkyl alcohols (Fig. 5). The size of the carbon chain had a minimum effect on the reaction conversion. Regardless of the alcohol (i.e., C1–C4), alkyl levulinate (AL) was always the main product.

Fig. 5
figure 5

Cs2.5H0.5PW12O40-catalyzed levulinic acid esterification with alkyl alcohols. Reaction conditions: levulinic acid (2 mmol), alkyl alcohol (8 mL), dodecane (internal standard), catalyst (1.2 mol%), temperature (393 K), time (6 h).

Conversely, the hysteric hindrance on the carbon bound to the hydroxyl group was crucial to the reaction conversion; the secondary alcohols were the less reactive. In all the reactions, alkyl levulinate selectivity was higher than 90%.

The reusability of the Cs2.5H0.5PW12O40 catalyst was assessed (Fig. 6). After the end of the reaction, the suspension containing the solid catalyst was centrifugated, and the solid catalyst was washed with ethyl alcohol, dried in an oven, weighted and reused.

Fig. 6
figure 6

Reuse of Cs2.5H0.5PW12O40 catalyst in the levulinic acid esterification with ethyl alcohols. Reaction conditions: levulinic acid (2 mmol), ethyl alcohol (8 mL), dodecane (internal standard), catalyst (1.2 mol%), temperature (393 K), time (6 h)

The conversion and selectivity of reactions remained almost constant after successive cycles of reuse of the Cs2.5H0.5PW12O40 catalyst. In all runs, the recovery rate was greater than 90%.

Conclusions

In this work, the activity of Cesium exchanged partially Keggin HPA salts was evaluated in the reactions of levulinic acid esterification with alkyl alcohols. Among the cesium salts of different HPAs containing 0.5 mol of H+ ions, the Cs2.5H0.5PW12O40 was the most active and selective catalyst. Investigations on the effect of H+ load per mol of phosphotungstate catalyst revealed that only the totally substituted salt (i.e., Cs3.0PW12O40) presented a conversion lower than 90%. The effect of Cs2.5H0.5PW12O40 load was also evaluated; 0.6 mol% of H+ ions was the lowest load that led to the highest conversion. The reactions achieved a satisfactory conversion only when carried at temperatures greater than 393 K and within a period equal to or higher than 6 h. When evaluated in esterification reactions of levulinic acid with primary alcohols (C1–C4 carbon atoms), the Cs2.5H0.5PW12O40 salt was also an efficient catalyst. However, the secondary alcohols were less reactive. The Cs2.5H0.5PW12O40 catalyst was successfully recovered and reused on levulinic acid esterification with ethyl alcohol without loss of activity.