A Comparative Study of Homogeneous and Heterogeneous Catalyzed Glycerol Acetylation with Acetic Acid: Activity, Selectivity, and Stability

Abstract

The esterification of glycerol with acetic acid to produce acetylglycerols was investigated both in the absence of a catalyst and using several commercial FAU zeolites with varied acid site concentrations and pore structures. Specifically, ultra-stable Y zeolite (H-USY, CBV720; Si/Al: 15), ammonium Y zeolite (NH₄-Y, CBV300, Si/Al: 2.6), and Na-Y zeolite (CBV100, Si/Al: 2.6) were tested to evaluate their catalytic performance in this reaction. In addition, two control catalysts were assessed under identical conditions: a commercial acidic ion-exchange resin (Amberlyst 15, A15) and a homogeneous acid catalyst (p-toluenesulfonic acid, PTSA). The catalytic performance of both the ultra-stable Y zeolite (CBV720) and Amberlyst 15 resin was comparable to the best results previously reported for solid catalysts. Furthermore, a comprehensive examination of acidity and accessibility was conducted to better understand their behavior. As expected, PTSA exhibited higher yields but also showed common drawbacks associated with homogeneous catalysts, such as corrosion and difficulties in separation. From an environmental perspective, and considering potential reuse cycles, the CBV720 zeolite delivered promising results during five consecutive tests. Its levels of conversion and selectivity were comparable to those obtained with Amberlyst 15, making it a viable alternative for future studies.

1. Introduction

Biodiesel is an alternative fuel that is highly desirable from a sustainability point of view because of its renewable origin [1]. It can be used in static engines and motor vehicles, either directly or blended with conventional diesel fuel. Its use offers several advantages. First, it has a lower environmental impact (lower CO₂ emissions because its production comprises a closed carbon cycle [2]). Second, biodiesel is produced from renewable raw materials like vegetable oils through transesterification of triglycerides with methanol over a basic catalyst. As a result of this process, glycerol can be obtained in significant yields, around 10% of the total volume of biodiesel produced [3]. Increased production and demand have led to a surplus of glycerol in the market, reducing its price and affecting the economics of biodiesel production. Considering also that the generated glycerol contains impurities (crude glycerol), such as methanol, soaps and water, the market value was US cent 8.81–19.84 per kilogram of crude glycerol in 2014 and almost close to US cent 59.52–90.39 per kilogram of pure glycerol [4]. The market values of glycerol up to 2024 remained within this range (around US cent 16.5 per kilogram of crude glycerol and close to US cent 90.0 per kilogram of pure glycerol) [5]. On the other hand, purification costs are relatively high, limiting its utilization in food additives, cosmetics, and surfactants [6]. However, glycerol’s availability represents an opportunity to obtain several value-added chemicals via highly selective catalytic conversion [3,7,8].

Given these challenges with crude glycerol utilization, researchers have explored new pathways to add value to this byproduct. An interesting option is the production of additives such as diacetylglycerols (1,2-DAG and 1,3-DAG) and triacetylglycerol (TAG). These compounds have several applications, including improving viscosity and low-temperature performance of petroleum fuels and biodiesel (such as pour point and cold filter plugging point). TAG may also serve as an antiknock additive for gasoline [7] and can enhance viscosity while meeting EN14214 and ASTM D6751 standards for flash point and oxidation stability [9,10,11]. Finally, a comprehensive review indicates that TAG’s use as a fuel additive is associated with environmental and economic benefits [12].

With respect to DAG and TAG production, two acetylating reactants have been studied: acetic acid (HAc) and acetic anhydride (Ac₂O). Ac₂O yields high selectivity for TAG [13], but presents higher risks and costs related to handling, transport, storage and equipment. Moreover, Ac₂O is a raw material for narcotic production. It is classified as contraband in most countries, which prohibits its use even for research purposes. It increases cost and limits supply; thus, acetic acid, being more readily available, is generally preferred [14].

Esterification of glycerol with acetic acid and other acids in the liquid phase proceeds through three consecutive reaction steps. Acetic acid (HAc) is successively added to glycerol (Gly), first forming monoacetylglycerols (1-MAG and 2-MAG), then diacetylglycerols (1,2-DAG and 1,3-DAG), and finally producing triacetylglycerol (TAG), with water generated as a by-product at each stage. Esterification studies with other alcohols demonstrate similar mechanisms on acidic solid catalysts, beginning with acid adsorption and followed by nucleophilic attack from the alcohol’s hydroxyl group on the carboxylic group [15].

In summary, these reactions are acid-catalyzed in both homogeneous and heterogeneous systems. Homogeneous catalysis with strong acids demonstrates high efficiency, following the Fischer esterification pathway. Notably, the reaction may proceed even without an external catalyst upon mere mixing of reactants; thus, experimental outcomes should always be benchmarked against control tests to account for potential autocatalysis by acetic acid. Since these reactions are reversible and controlled by thermodynamic equilibrium, developing catalysts that achieve high glycerol conversions and favor production of highly acetylated esters (DAG and TAG) remains challenging.

Currently, this reaction is performed industrially without catalysts or with homogeneous liquid acid catalysts such as p-toluene sulfonic acid (PTSA), H₃PO₄, H₂SO₄, and HCl [16]. Homogeneous acids cause issues like reactor corrosion, harmful waste generation, and catalyst loss due to incomplete separation. To mitigate these issues, various heterogeneous acid catalysts have been explored, including alkyl-sulfonic acid functionalized SBA-15, commercial resins, functionalized mesostructured solids, and carbon-based catalysts [17,18,19,20,21,22,23,24,25].

Recent comprehensive reviews by Hameed et al. [14] and Pérès et al. [26] have assessed the state of catalytic acetylation of glycerol. These works emphasize the ongoing need for new catalytic solids and more detailed studies, due to challenges such as side-product formation, issues arising from poorly structured pore channels, catalyst deactivation or thermal degradation, and the leaching of functional groups. The performances of the main catalysts are presented in Table 1

Beyond traditional acid catalysts, zeolitic materials have also been explored due to their unique pore structure. Some studies have investigated glycerol acetylation using both commercial and modified zeolites such as H-ZSM-5, H-BEA, and H-Y. These studies show lower catalytic activity compared to other acidic solids due to no adequate zeolite selection from the point of view of active site accessibility and hydrophilic character [18,27,28,29,30,31]. Despite this, H-Y zeolites have been widely employed for esterification with various alcohols and acids. For instance, H-Y zeolite catalyst CBV780 (Si/Al:40) showed high efficiency in the esterification of oleic acid with methanol, achieving a 98% conversion rate and promising recyclability [32]. Additionally, it was found that low-silica Y zeolites (silicon-to-aluminum ratios of 2 and 5) are susceptible to dealumination by lactic acid under reaction conditions, rendering them unsuitable as catalyst supports [33]. In addition, other studies have examined H-USY zeolite for esterifying levulinic acid with different alcohols [34,35], fatty acids with methanol using CBV300, CBV720, and H-Y zeolites [36,37,38], and acetic acid with n-butanol using heteropolyacid-modified CBV300 zeolite [39].

Table 1. Solid Catalysts performances in the glycerol acetylation with acetic acid.

Catalyst	HAc/Gly a	Temp (°C)	Time (h)	X (%)	S (%) b	Recycles	Ref
A15	9/1	110	4.5	97.1	92.2	Not Performed	[25]
H-USY				78.4	26.2	Not Performed
H-ZSM-5				85.6	33.4	Not Performed
A15	6/1	105	4.0	98.6	25.0 *	Not Performed	[18]
H-Y	6/1	105	4.0	80.9	3.2 *	Not Performed
H-ZSM-5	6/1	105	4.0	75.7	2.5 *	Not Performed
Sulf-SBA-15	6/1	120	4.5	100	70.0	Not Performed	[40]
H-ZSM-5 AT	6/1	120		91.0	52.0	Not Performed
PMo3_Na-USY	15/1	120	3	68.0	61.0	5	[23]
PW2_AC	15/1	120	3	86.0	74.0	4	[24]

^a Molar Ratio, ^b Selectivity to D.A and T.A, * Selectivity to T.A.

Table 1. Solid Catalysts performances in the glycerol acetylation with acetic acid.

Catalyst	HAc/Gly a	Temp (°C)	Time (h)	X (%)	S (%) b	Recycles	Ref
A15	9/1	110	4.5	97.1	92.2	Not Performed	[25]
H-USY				78.4	26.2	Not Performed
H-ZSM-5				85.6	33.4	Not Performed
A15	6/1	105	4.0	98.6	25.0 *	Not Performed	[18]
H-Y	6/1	105	4.0	80.9	3.2 *	Not Performed
H-ZSM-5	6/1	105	4.0	75.7	2.5 *	Not Performed
Sulf-SBA-15	6/1	120	4.5	100	70.0	Not Performed	[40]
H-ZSM-5 AT	6/1	120		91.0	52.0	Not Performed
PMo3_Na-USY	15/1	120	3	68.0	61.0	5	[23]
PW2_AC	15/1	120	3	86.0	74.0	4	[24]

^a Molar Ratio, ^b Selectivity to D.A and T.A, * Selectivity to T.A.

Moreover, zeolites have also been studied for other value-added transformations of glycerol. Research has demonstrated the use of H-Beta and H-Y zeolites in processes such as glycerol etherification and acetalization [40,41]. These findings highlight the versatility of zeolitic materials as catalysts in the conversion of glycerol into higher-value chemicals. However, it should be noted that glycerol acetylation has not been specifically examined in these cases. Among the variety of catalytic solids explored, Faujasite zeolites remain less investigated with respect to glycerol acetylation.

Previous works by the authors [42] demonstrated that the conversion of glycerol to MAG can be catalyzed by MFI zeolites, but subsequent transformation to DAG and TAG is hindered by the porous structure, obtaining results no better than a blank test. Development of mesoporosity by alkaline treatment over these catalysts [43] affected the availability of acid sites required by the reaction, reporting no improvements in the activity. Then, H-ZSM-5 exhibited poor performance owing to its porous structure. Further tests with mesoporous materials [42,44] showed that the intrinsic activity of the sites of sulfonic silica is comparable to sulfuric acid, though the increment in the load of sites led to steric hindrances, which were less pronounced in well-ordered SBA-15 catalysts. Considering their pore size and stability, as well as their widespread use in industrial catalytic processes, FAU zeolites are good candidates for this reaction. Although CBV720 zeolite displays moderate acidity, it offers suitable porosity and hydrophobicity. Our aim is to enhance hydrophobicity and accessibility to active sites compared with zeolites previously studied [18,27,31]. Finally, A15, which has been identified as one of the most active catalysts in the literature, was included in the study as a reference catalyst and evaluated under the same reaction conditions as the zeolites

Despite prior research, a comprehensive comparison of homogeneous and heterogeneous catalysts for this reaction remains unavailable. In the present contribution, two Y zeolites were examined, and their catalytic performance was compared to that of a commercial resin (A15). The reusability of these catalysts was assessed by conducting two consecutive recycles, with one selected Y-zeolite catalyst subjected to five additional recycle tests to establish its stability. Furthermore, p-toluene sulfonic acid, a widely used industrial homogeneous catalyst for this reaction, was tested as a control, with its performance directly contrasted with those of the heterogeneous catalysts.

2. Materials and Methods

2.1. Materials

Several commercial zeolites supplied by Zeolyst^® (Kansas City, KS, USA) were selected: Protonic ultra-stable Y zeolite, H-USY (CBV720; Si/Al: 15); Ammonium Y zeolite, NH₄-Y (CBV300, Si/Al: 2.6) and Sodium Y zeolite, Na-Y (CBV100, Si/Al: 2.6). Amberlyst-15 (A15), a commercial resin obtained from Sigma Aldrich (Sofia, Bulgaria), was also tested as a benchmark for comparison. For the catalytic tests the following compounds from Sigma-Aldrich were used: glycerol (>99.0%), glacial acetic acid (>99.8%) and p-toluene sulfonic acid (PTSA). Before its catalytic evaluation, NH₄-Y zeolite was calcined at 823 K for 5 h (2 °C min⁻¹) in a muffle furnace overnight to obtain the corresponding H-Y zeolite.

2.2. Characterization Techniques

A comprehensive set of analytical methods was employed to characterize the physicochemical properties of the catalysts. These included N₂ adsorption–desorption isotherms at 77 K, transmission electron microscopy (TEM) combined with energy dispersive X-ray spectroscopy (EDX), scanning electron microscopy (SEM), and X-ray diffraction (XRD) analyses. The acidity of the materials was assessed via temperature-programmed desorption (TPD) of pyridine and collidine, as well as Fourier-transform infrared spectroscopy (FTIR).

The textural properties of the solids were determined from N₂ adsorption–desorption isotherms at 77 K obtained at the technical services of the University of Alicante (SSTTI-UA) using a Quantachrome AUTOSORB-6 equipment (Anton Paar, Graz, Austria). Structural characterization of the zeolites was conducted with powder XRD on a SEIFERT 2002 instrument (Mannheim, Germany) using CuKα (λ = 1.5418 Å) radiation. The study was carried out using wide-angle XRD in the 2.5 < 2θ < 50 range, using a scanning velocity of 1° min⁻¹. The powders were mixed with graphite as an internal standard to determine the crystallinity of the samples. The intensity of the peak at 2θ = 26.5° of the graphite was used to normalize the diffraction patterns, and relative crystallinity was calculated using the peak at 2θ = 16° of the zeolite.

The morphology and microstructure of the samples were examined by TEM at the SECEGRIN facility of CCT-CONICET Santa Fe, and by SEM at SSTTI-UA. TEM imagery was obtained with a JEOL JEM-2100 Plus microscope (Akishima, Japan) equipped with a LaB₆ gun. Samples were suspended in ethanol and then two drops were deposited over a copper-grid. Bright-field images were obtained under a voltage of 200 kV. Image processing was performed with Gatan Microscopy Suite version 3.6.0 under a free license. For further analysis of calcined zeolite cross-sections, samples were embedded in Spurr resin and sectioned into 80 nm slices using an RMC MTXL ultramicrotome (Boeckeler, Tucson, AZ, USA). Then, TEM images were captured with a JEOL JEM-1400 Plus transmission electron microscope (Akishima, Japan) operating at 120 kV, with a resolution of 0.38 nm between points and a line resolution of 0.20 nm and equipped with a GATAN ORIUS image acquisition camera (Las Positas Blvd. Pleasanton, CA, USA). SEM images were obtained using a Hitachi S-3000N microscope equipped with a scintillator-photomultiplier secondary electron detector (3.5 nm resolution), a semiconductor backscattered electron detector (5 nm resolution), and a Bruker XFlash 3001 X-ray detector (Billerica, MA, USA) for microanalysis.

To evaluate the quantity and strength of acid sites, pyridine TPD measurements were performed. Collidine (2,4,6-trimethylpyridine), a bulkier base, was used in parallel to determine the accessibility of acid sites in the zeolites. The nature of acid sites was further investigated by collecting FTIR spectra (INCAPE-CONICET) following pyridine and collidine saturation, using a Shimadzu Prestige 21 spectrometer (Kyoto, Japan). Additional details regarding experimental procedures are provided in Appendix B.

2.3. Catalytic Tests

All catalytic reactions were performed at atmospheric pressure within a 250 mL three-neck flask, equipped with stirring and heated using an oil bath controlled by a thermocouple and a PID controller. The reaction temperature was maintained at 393 K. To ensure the retention of water and acetic acid vapors, a condensing column was utilized. The experimental procedure involved heating the mixture of glycerol and solid catalyst to the desired temperature, followed by injection of acetic acid into the system. Three different acetic acid to glycerol molar ratios were explored: 12/1, 6/1, and 3/1. The zeolites were calcined at 823 K for 5 h prior to use, and the dry A15 was kept in an oven at 353 K to avoid moisture adsorption.

Tests were conducted under two main conditions: (i) maintaining a constant solid catalyst to glycerol mass ratio of 2 g catalyst per 100 g glycerol, and (ii) using a fixed acid sites per gram of glycerol feed ratio (9.5 × 10⁻² mmol/g glycerol), applicable for A15, CBV720 zeolite, and PTSA. Due to the diffusional limitations of the microporous CBV300 zeolite, it was tested at approximately five times this feed ratio. The reuse of both zeolites and the resin was studied in different catalytic cycles. After each cycle, the catalysts were separated by centrifugation and regenerated by thoroughly washing them with ethanol and then air dried at 105 C for 12 h.

To determine the glycerol conversion and products selectivity, samples were collected every 30 min, filtered and analyzed by gas chromatography (Shimadzu GC-2010, Kyoto, Japan) equipped with a flame ionization detector (FID), AOC-20i autosampler (Shimadzu, Kyoto, Japan), and a BP20 column (30 m × 0.25 mm × 0.25 μm, Sprockhövel, Germany).

Glycerol conversion (X) and product selectivity (S_α) were calculated with the following equations:

$$\mathrm{X}={\displaystyle \frac{\mathrm{a}\mathrm{m}\mathrm{o}\mathrm{u}\mathrm{n}\mathrm{t} \mathrm{o}\mathrm{f} \mathrm{c}\mathrm{o}\mathrm{n}\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{t}\mathrm{e}\mathrm{d} \mathrm{g}\mathrm{l}\mathrm{y}\mathrm{c}\mathrm{e}\mathrm{r}\mathrm{o}\mathrm{l}}{\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l} \mathrm{a}\mathrm{m}\mathrm{o}\mathrm{u}\mathrm{n}\mathrm{t} \mathrm{o}\mathrm{f} \mathrm{f}\mathrm{e}\mathrm{d} \mathrm{g}\mathrm{l}\mathrm{y}\mathrm{c}\mathrm{e}\mathrm{r}\mathrm{o}\mathrm{l}}}$$

$${\mathrm{S}}_{\mathsf{\alpha }}={\displaystyle \frac{\mathrm{a}\mathrm{m}\mathrm{o}\mathrm{u}\mathrm{n}\mathrm{t} \mathrm{o}\mathrm{f} \mathrm{g}\mathrm{l}\mathrm{y}\mathrm{c}\mathrm{e}\mathrm{r}\mathrm{o}\mathrm{l} \mathrm{c}\mathrm{o}\mathrm{n}\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{t}\mathrm{e}\mathrm{d} \mathrm{t}\mathrm{o} \mathrm{p}\mathrm{r}\mathrm{o}\mathrm{d}\mathrm{u}\mathrm{c}\mathrm{t} \mathsf{\alpha }}{\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l} \mathrm{a}\mathrm{m}\mathrm{o}\mathrm{u}\mathrm{n}\mathrm{t} \mathrm{o}\mathrm{f} \mathrm{c}\mathrm{o}\mathrm{n}\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{t}\mathrm{e}\mathrm{d} \mathrm{g}\mathrm{l}\mathrm{y}\mathrm{c}\mathrm{e}\mathrm{r}\mathrm{o}\mathrm{l}}}$$

The specific activity of the acid sites for glycerol conversion under the defined reaction conditions was assessed using the turn-over frequency (TOF, h⁻¹). TOF was calculated by dividing the number of moles of glycerol converted per mole of catalyst (acid sites) and the reaction time.

$$\mathrm{T}\mathrm{O}\mathrm{F} [{\mathrm{h}}^{-1}]={\displaystyle \frac{\mathrm{a}\mathrm{m}\mathrm{o}\mathrm{u}\mathrm{n}\mathrm{t} \mathrm{o}\mathrm{f} \mathrm{c}\mathrm{o}\mathrm{n}\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{t}\mathrm{e}\mathrm{d} \mathrm{g}\mathrm{l}\mathrm{y}\mathrm{c}\mathrm{e}\mathrm{r}\mathrm{o}\mathrm{l} \left[\mathrm{m}\mathrm{o}\mathrm{l}\right]}{\mathrm{m} \mathrm{c}\mathrm{a}\mathrm{t} \left[\mathrm{g}\right] \times \mathrm{t} \left[\mathrm{h}\right] \times \mathrm{a}\mathrm{c}\mathrm{i}\mathrm{d} \mathrm{s}\mathrm{i}\mathrm{t}\mathrm{e}\mathrm{s} [{\mathrm{m}\mathrm{o}\mathrm{l} \mathrm{H}}^{+} {\mathrm{g}\mathrm{c}\mathrm{a}\mathrm{t}}^{-1}]}}$$

Catalysts classified as “used” were those subjected to reaction for 24 h. Following the reaction, these catalysts were filtered, centrifuged, washed, and dried. For zeolite solids, an additional gradual calcination at 823 K for 5 h (1° min⁻¹) was performed to complete the post-reaction treatment.

3. Results

3.1. Structure Characterization

Figure 1 presents the N₂ adsorption–desorption isotherms at 77 K for the CBV300 and CBV720 zeolites, as well as the A15 resin. Using these isotherms, the corresponding textural parameters were calculated and are summarized in

Table 2. Textural and acidity results.

Sample	Fresh Catalyst				After Reaction
	Vmicro (cm3 g−1)	Vmeso (cm3 g−1)	Dp (nm)	Total Acid Sites (mmol g−1)	Vmicro (cm3 g−1)	Vmeso (cm3 g−1)	Dp (nm)	Total Acid Sites (mmol g−1)
A15	-	-	30	4.7 *	-	-	30	n.d
CBV300	0.32	0.06	1.17	0.94 **	0.34	0.09	1.17	0.62
CBV720	0.37	0.24	1.22	0.36 **	0.32	0.28	1.22	0.32

* Supplier data. ** Pyridine –TPD measurements.

. The fresh zeolites (with solid and empty red circles representing adsorption and desorption branches, respectively) display Type I isotherms which are characterized by high nitrogen uptake at low relative pressures (P/P° < 0.2) characteristic of microporous materials. The absence of hysteresis cycle in CBV300 isotherm (Figure 1A) indicates the low number of defects of this zeolite and the absence of mesoporosity [45,46]. In the case of the CBV720 sample (Figure 1B), a significant increase in N₂ adsorption is observed at relative pressures between 0.80 and 0.95, indicating the existence of large mesopores alongside the micropores. These features result from the steaming process performed by the supplier [47]. At even higher relative pressures (above 0.95), both materials exhibit additional adsorption, which is attributed to interparticle condensation. As anticipated, the A15 resin’s isotherm (Figure 1C) demonstrates the typical profile of a macroporous material.

When examining the zeolites after catalytic use (solid and empty black squares for adsorption and desorption branches, respectively), a distinct change in behavior emerges. The used CBV720 zeolite retains an isotherm profile very similar to that of the fresh material, except for a slight reduction in micropore volume (see Table 2). In comparison, the CBV300 zeolite exhibits a modest increase in N₂ adsorption at P/P° > 0.1. These findings indicate that under reaction conditions, (i) a small blocking of the microporosity occurs in the CBV720 zeolite, which is not totally recovered after the calcination step; and (ii) the CBV300 zeolite develops modified porosity, likely caused by the acetic acid and water present in the reaction mixture—a phenomenon previously reported for low-Si/Al-ratio zeolites exposed to acidic solutions [32].

XRD patterns for the CBV300 and CBV720 zeolites, both before and after the reaction, are shown in Figure 2. For each sample, graphite was added as an internal standard, allowing normalization of the diffraction data by referencing the most intense graphite peak [(0 0 2) at 26.55° 2θ]. The XRD results confirm that all samples exhibit the faujasite (FAU) framework, with no other crystalline phases detected. Importantly, this FAU structure persisted throughout the catalytic testing process. Notably, the CBV720 zeolite demonstrated consistent crystallinity before and after the reaction, showing no significant changes. In contrast, the CBV300 zeolite experienced a decrease in crystallinity of approximately 10% following the reaction. This observation suggests that the zeolite with the higher aluminum content in its framework was more susceptible to alteration under the tested reaction conditions. This is consistent with the N₂ isotherm results.

To further investigate the morphology of the samples, transmission electron microscopy (TEM) images for both zeolites and scanning electron microscopy (SEM) images for the A15 resin were obtained. First, an exploratory analysis was performed (refer to Appendix A, Figure A1). The TEM micrograph of the H-USY zeolite (Figure A1B) revealed the presence of mesopores generated by the steaming process, consistent with findings from the N₂ adsorption isotherm. In contrast, these mesopores were absent in the CBV300 zeolite, as shown in Figure A1A. SEM analysis of the A15 resin revealed that it is composed of spherical particles ranging from 300 to 800 μm in diameter (Figure A1C).

To better distinguish morphological differences between the two zeolites, TEM micrographs of ultramicrotomed samples were also taken and are presented in Figure 3.

Additional TEM analysis was conducted on zeolites after their use in reaction. The results indicated that CBV720 retained its morphological features (see rows III and IV in Figure 3), whereas the used CBV300 zeolite displayed the presence of external aggregates, suggesting the potential formation of segregated extra-framework aluminum species (see rows I and II in Figure 3). This hypothesis was supported by Energy Dispersive X-ray (EDX) analysis at both general and specific locations on the samples. For CBV720, EDX quantification revealed an average Si/Al ratio of approximately 20, with a homogeneous distribution observed for both the calcined and the used material. In contrast, the calcined and used CBV300 samples showed average Si/Al ratios of about 2.58 and 3.04, respectively, with a more heterogeneous distribution. These results indicate a decrease in bulk aluminum content for CBV300. Furthermore, EDX analysis of the surface aggregates in the used CBV300 sample showed Si/Al ratios between 0.23 and 0.70, confirming the preferential segregation of aluminum in this material.

All mentioned TEM micrographs, accompanied by the EDX maps of Si and Al and spectra of selected regions, are shown on Figure A2 in Appendix A.

In conclusion, the presence of excess acetic acid and water generated during the reaction can interact with the aluminum atoms in the CBV300 zeolite framework. This interaction may induce a certain degree of dealumination, which in turn alters the porosity of the material—as observed in both TEM and N₂ adsorption measurements—and results in a decreased degree of crystallinity as observed in XRD spectra.

3.2. Acidity Characterization

3.2.1. Temperature-Programmed Desorption (TPD) of Pyridine and Collidine

The acidity profiles of both zeolites, before and after reaction, were investigated using temperature-programmed desorption (TPD) of pyridine and collidine. This analysis is summarized in

Table 3. Amount of acid sites determined by TPD with different probe molecules, for zeolites before and after reaction.

		Pyridine (mmol g−1)	Collidine (mmol g−1)	Col/Py (%)
CBV300	Calcined	0.938	0.580	61.8
	Used	0.619	0.566	91.4
	Used/Calcined (%)	66.0	97.6	-
CBV720	Calcined	0.361	0.250	69.3
	Used	0.324	0.259	79.9
	Used/Calcined (%)	89.8	~100	-

and Figure 4. The FAU structural model features supercages with a free diameter of 1.12 nm, each interconnected by four windows with a free diameter of 0.730 nm [48]. Given the kinetic diameters of pyridine (0.533 nm) and collidine (0.740 nm), these molecules exhibit varying degrees of accessibility to the microporous acid sites. For the commercial CBV300 and CBV720 solids, approximately 60% and 70% of acid sites, respectively, were accessible to collidine (as shown in Table 2, Col/Py (%)).

After reaction, CBV720 maintained nearly 90% of acid sites accessible to pyridine (Table 3 and Figure 4C,D, continuous line), whereas CBV300 retained only 66% (Table 3 and Figure 4A,B, continuous line). This reduction aligns with the modifications in porosity and crystallinity previously observed. Importantly, the quantity of acid sites accessible to collidine remained largely unchanged for both zeolites, indicating that the acid sites lost during reaction were predominantly located in micropores inaccessible to collidine. Further examination of the TPD profiles revealed that CBV300, following reaction, preferentially lost acid sites of medium and high strength (T > 673 K), most of which were situated in positions accessible to pyridine. In contrast, CBV720 preserved its pyridine desorption profile with minimal changes after reaction.

3.2.2. Pyridine FTIR Studies

OH Region

The FTIR spectra of CBV300 and CBV720 zeolites in the OH stretching region are illustrated in Figure 5. For CBV720, the spectrum displays three distinct and intense OH stretching bands at 3744, 3628, and 3566 cm⁻¹. Both lower-frequency bands show components at their lower-frequency sides. This pattern is characteristic of ultrastable Y-type (USY) zeolites [49,50], while the other two bands correspond to “structural” hydroxyl groups: high-frequency (HF) OH groups in the supercages (3628 cm⁻¹) and low-frequency (LF) OH groups in the sodalite cages (3566 cm⁻¹). The low-frequency shoulders near 3600 and 3550 cm⁻¹ are associated with HF and LF species interacting with residual extra-framework (EF) material [51]. The CBV300 sample exhibits a much weaker silanol band at 3742 cm⁻¹, along with a split band at 3689 and 3678 cm⁻¹ in the region typically assigned to OH groups on extra-framework material. Finally, two bands are found at 3636 and 3551 cm⁻¹, possibly due to two components, in the region of zeolitic hydroxyl groups. The Na-Y sample displayed extremely weak bands at 3741 and 3684 cm⁻¹, suggesting that only a minimal amount of external silanol groups and OHs on extra-framework species is present. Bands due to bridging OHs in the supercage (HF) and sodalite cage (LF) are absent, indicating complete cation exchange by sodium ions.

Upon comparing the catalysts before and after use, LF bands disappear entirely, and HF bands decrease in intensity and shift to lower frequencies. Conversely, silanol and extra-framework aluminum (EFAL) bands intensify. This suggests that aluminum is extracted from the framework, forming defects and EFAL species, which interact with the remaining sites and cause the observed shifts in absorption bands. CBV300 demonstrates a greater reduction in HF and LF bands, consistent with its larger loss of crystallinity, increased mesoporosity, and TEM-EDX findings.

Pyridine Adsorbed Region

Figure 6 shows the FTIR surface acidity study of the CBV300 and CBV720 samples, by using pyridine as a probe molecule. Spectra were recorded after three outgassing steps at 423, 573 and 673 K. The pyridine species present after the first step were considered to correspond to the total amount of acid sites. After the second step, the remaining pyridine was assigned to “medium” and “strong” sites, having desorbed the pyridine from the “weak” sites. Finally, sites corresponding to pyridine species still found after outgassing at 673 K are assumed to be the “strong” fraction. The number of sites corresponding to each fraction was determined by subtracting the associated values (

Table 4. Acid sites characterization: Lewis and Brønsted sites concentrations determined by FTIR analyses of adsorbed pyridine.

	Strength	Br/Lw	Br (μmol g−1)	Lw (μmol g−1)
CBV300	Weak	2.192	344	157
	Medium	7.391	396	54
	Strong	0.955	166	174
CBV720	Weak	3.539	43	12
	Medium	5.764	78	14
	Strong	3.751	142	38
used CBV300	Weak	1.243	337	271
	Medium	1.294	85	66
	Strong	0.953	91	95
used CBV720	Weak	1.068	126	118
	Medium	1.330	93	70
	Strong	0.450	26	59

). This method ensured accurate representation of the distribution across all fractions and provided a clear quantitative basis for subsequent analysis. At each step, quantification was carried out according to the procedure described by Ivanova et al. [52].

The spectra feature bands at 1633–1631 cm⁻¹ (vibrational mode 8a) and 1545 cm⁻¹ (mode 19a) [53], assigned to pyridinium ions formed by protonation of pyridine on Brønsted acid sites. Bands at 1622–1624 and 1452–1455 cm⁻¹ correspond to very strong Lewis acid sites, typical of alumina and silica-alumina materials [54,55]. Comparing the two protonic zeolites, the relative intensities of pyridinium and molecular pyridine bands reveal that Brønsted sites are more prominent in H-USY (CBV720) than in H-Y (CBV300).

The Na-Y (CBV100, not depicted) sample contains only Lewis acid sites, in greater quantity than the protonic zeolites. While these Lewis sites are considered “strong,” they are weaker than those in the protonic zeolites and are mainly associated with Na⁺ and Al^3⁺ cations.

In Table 4, it is observed that CBV300 has a similar number of strong Lewis and Brønsted sites, with Brønsted sites predominating in medium and weak strengths—over 80% combined—while the strong fraction dominates among Lewis sites. CBV720 exhibits a higher proportion of strong sites for both Brønsted and Lewis acidity, with Brønsted sites prevailing across all strengths (Br/Lw ratios 3.54–3.76). The overall trend for total (Lewis + Brønsted) strong acid site density is CVB300 > CVB720 > CVB100. Among the protonic zeolites, CBV300 has more Brønsted and Lewis acid sites than CBV720, which reflects its higher Al content, but both have similar amounts of strong Brønsted sites.

The presence of Lewis sites on CBV300 zeolite is likely due to extra-framework material, as the 3698 and 3678 cm⁻¹ bands indicate OH groups on EFAL (Figure 5). For CBV720, the existence of Lewis acid sites is not straightforward, given the low framework Al content and the lack of EFAL in the OH region, though active Lewis sites have been previously reported in USY [46,56,57]. Weak Brønsted sites are mostly silanols. Quantification results summarized in Table 4 indicate that the Br/Lw ratio from pyridine-FTIR is higher for CBV720 than CBV300, and about 54% of Brønsted sites in CBV720 are strong, compared to just 18% in CBV300. In addition, acid sites accessible to collidine are around 80% in CBV720 and 30% in CBV300.

After reaction, both catalysts show a decrease in the total number of acid sites, particularly strong sites, with a corresponding increase in weak sites. The Br/Lw ratio drops to approximately 1.2, indicating a more substantial loss of Brønsted than Lewis sites, consistent with the changes in OH groups observed earlier. Notably, CBV720 loses strong Brønsted sites but gains a slight increase in strong Lewis sites, suggesting some transformation of acid sites and the possible formation of EFAL or defects. Overall, Brønsted sites are converted into Lewis sites, and both types are weakened and diminished.

3.3. Reactions Catalyzed by Acetic Acid (Autocatalytic)

A series of experiments was conducted to evaluate conversion and selectivity in the absence of an added catalyst, relying solely on acetic acid to provide the necessary protons for acetylation. Three tests were performed using molar HAc/Gly ratios of 3/1, 6/1, and 12/1. These experiments measured glycerol conversion (see Figure A3A and

Table A1. Autocatalytic Reactions: Glycerol conversion.

Reagents Molar Ratio HAc/Gly	X (%) (5 h)	X (%) (24 h)
3/1	84.6	91.2
6/1	91.0	95.5
12/1	98.7	100.0

in Appendix A), as well as selectivity toward MAG, DAG, and TAG (Figure A3B–D). For all reactions, the same volume was maintained.

Initially, reaction rates appeared similar across all ratios. However, after several minutes, the test with the highest acetic acid content (12/1, represented by the blue curve) exhibited an accelerated reaction rate, with acetic acid acting as the catalyst. After 24 h, complete conversion (100%) was achieved only in the 12/1 test; nevertheless, high conversions were also obtained with the other ratios, reaching 95.5% for 6/1 and 91.2% for 3/1 (see Table A1 and Figure A2a).

In terms of selectivity toward MAG, DAG, and TAG, the highest selectivity to TAG was observed in the reaction with the greatest amount of acetic acid, consistent with previous literature [8,15]. Selectivity data at two specific glycerol conversions (10% and 80%), as well as final selectivity values, are summarized in

Table A2. Autocatalytic Reactions: DAG and TAG selectivity.

Reagents Molar Ratio HAc/Gly	X: 10 %		X: 80 %		Final
	Time (min) *	SDAG + STAG (%)	Time (min) *	STAG + SDAG (%)	SDAG + STAG (%)
3/1	13	4.0 [0]	240	39.3 [3.45]	61.3 [10.8]
6/1	10	6.0 [0]	240	39.6 [5.10]	78.8 [24.1]
12/1	4	4.0 [0]	120	45.6 [4.76]	92.3 [45.9]

* Time (in minutes) to get this glycerol conversion. [ ] TAG selectivity.

and Figure 7. The time-course evolution of selectivity is illustrated in Figure A3B–D. At low conversion (10%), product selectivity showed little variation between tests (Figure 7A). At higher conversion (80%), selectivity for TAG and DAG increased in the reaction with the lowest glycerol content (12/1 HAc/Gly molar ratio), as depicted in Figure 7B.

3.4. Reactions Catalyzed by Solid Acids

Following the exploration of autocatalytic reactions, it was observed that while the catalytic results achieved with a 12/1 acetic acid to glycerol molar ratio were slightly superior, such a high molar ratio is not applicable for the industrial process not only by reagent costs but due to the difficulty of the separation of the unreacted HAc. Therefore, an increased acetic acid requirement not only would affect the overall economics but also possess challenges for practical application. To address the cost and efficiency concerns, the catalytic behavior of three different solid acids was evaluated by using an acetic acid to glycerol molar ratio of 6/1, aiming to optimize the process by balancing catalytic performance and economic and operational feasibility.

Each catalyst was tested at a fixed dosage, corresponding to 2 wt.% of the initial glycerol mass. Results from these experiments are illustrated in Figure A4, Appendix A.

Among the catalysts, A15 demonstrated significantly higher activity than both zeolites, attributed to its higher amount of acid sites. The catalytic performance of the zeolites was comparable to that observed in the autocatalytic tests, which is likely due to their lower number of acid sites and the relatively small catalyst-to-glycerol mass ratio. Notably, CBV720, despite being less acidic than CBV300 ([H⁺] CBV720 ≈ 1/5 [H⁺] CBV300), achieved slightly superior performance. This enhancement is most likely related to the presence of mesopores, which facilitate diffusion and minimize transport limitations.

To assess whether the acid sites of the catalysts were equally effective, experiments were conducted with A15 and CBV720, ensuring the same number of acid sites per gram of initial glycerol. For CBV300, which showed reduced activity compared to CBV720, the same mass was used as for CBV720, resulting in a fivefold increase in the total concentration of acid sites. This allowed for evaluation of whether an excess of acid sites could compensate for limited accessibility, potentially leading to similar activity levels.

When the number of acid sites was standardized, the performance of CBV720 and A15 was nearly indistinguishable (see Figure 8A), indicating that the diffusivity of reactants and products was comparable in both catalysts, despite the smaller pore size in the zeolite versus the resin. In contrast, CBV300 still exhibited lower activity, even with a fivefold increase in acid site concentration, suggesting that not all its acid sites were accessible to the reactants.

The selectivity profiles for all catalysts revealed a typical sequence of consecutive reactions (Figure 8B–D). Initially, MAG was the predominant product, but as the reaction progressed, its selectivity decreased while that of DAG and TAG increased. After DAG selectivity reached its maximum, the rate of TAG formation slowed. In the final stages, all three catalysts converged to exhibit similar selectivity profiles, indicating that the ultimate product distribution was governed by chemical equilibrium.

To further investigate the contribution of the zeolitic support relative to its Brønsted acidity, a sodium zeolite (CBV100) was used as a control. The catalytic activity of CBV100 was found to be nearly identical to that observed in the autocatalytic reaction. This result indicates that the advancements seen with acid zeolites, specifically CBV720 and CBV300, are attributable exclusively to their Brønsted acidity. While CBV100 possessed a significant number of Lewis acid sites, these did not notably influence the reaction outcome.

The reaction was also conducted using p-toluene sulfonic acid as a homogeneous liquid catalyst, at a dosage of 9.5 × 10⁻² mmol H⁺ per gram of glycerol. This allowed a direct comparison between heterogeneous and homogeneous catalysis. The homogeneous catalyst demonstrated both higher initial activity and greater turnover frequency (TOF) at 30 min compared to the heterogeneous systems. Nevertheless, the final conversion and product selectivity observed with the homogeneous catalyst were comparable to those achieved with the A15 and CBV720 catalysts, as detailed in

Table A3. TOF and Selectivity. HAc/Gly molar ratio 6/1.

	TOF (h−1) (t = 30 min)	X (%) (t = 30 min)	X (%) (t = 24 h)	STAG (%) (t = 24 h)	SDAG + STAG (%) (t = 24 h)
A15	120.5	52.00	100	29.42	82.50
CBV720	113.3	48.70	98.50	27.71	80.90
PTSA	231.5	98.30	100	31.62	84.90

. Beyond catalytic performance, heterogeneous reactions offer notable advantages, particularly in terms of environmental impact and ease of catalyst separation.

3.5. Comparison Between Heterogeneous and Homogeneous Catalysts

To provide a thorough comparison, glycerol conversion was evaluated at different concentrations of p-toluenesulfonic acid (PTSA) used as the catalyst. Measurements were taken at 10, 30, and 60 min, with PTSA concentrations ranging from 0 to 0.024 mol H⁺ L⁻¹. These results enabled the construction of calibration curves to correlate the activity of various catalytic solids with the equivalent concentration of acidic PTSA sites needed to achieve similar conversion levels. The calibration curve at 10 min was selected for its high degree of linearity (see Figure 9A).

For example, CBV300, at an experimental acid site concentration of 0.102 mol L⁻¹, achieved a 24.5% glycerol conversion at 10 min. According to Figure 9B, this same conversion could be reached with just 0.00082 mol L⁻¹ of PTSA. Applying this analysis to CBV720 and A15 (each with an experimental proton concentration of approximately 0.0204 mol L⁻¹), the observed activity corresponded to 0.0014 mol L⁻¹ of PTSA, which was significantly higher than CBV300. Notably, this value is twenty times smaller than the actual concentration of acidic sites present in the solids. In summary, the activity level demonstrated by the solids could be matched using PTSA with only a twentieth of the proton concentration found in the solid catalysts. This observation highlights that, although the solid catalysts exhibited considerable activity, their performance remained considerably lower than that of the homogeneous catalyst. Nonetheless, this drawback can be compensated by the fact that solid catalysts can be easily separated from the reaction media and then reused.

The reusability of A15, CBV720, and CBV300 was evaluated and is illustrated in Figure 10 (A, B, and C, respectively). All three solid catalysts demonstrated comparable trends in conversion and selectivity between their initial use and subsequent reuse. While a slight decrease in catalytic activity was noted after recycling, the differences remained minimal. Based on the results from XRD analysis, surface area measurements, and acidity studies, this minor decline in performance can be attributed to the loss of a small fraction of the acid sites on the solid catalysts.

Finally, to assess CBV720 catalyst stability, five consecutive reaction tests were performed. Figure 11A,B show results at 5 and 24 h. Throughout the five catalytic cycles, both glycerol conversion and product selectivities were consistently preserved, indicating satisfactory catalyst performance over repeated use. Minor decreases in the quantity of the solid catalyst were observed. These diminutions were attributed to issues arising during the sampling process, rather than catalyst degradation. During the consecutive tests, the amounts of acetic acid and glycerol were modified as needed to ensure that the acid site-to-glycerol mass ratio remained constant. This adjustment was essential for maintaining uniform reaction conditions across all cycles

4. Discussion

The results demonstrate that glycerol acetylation with acetic acid, even without a catalyst, can achieve substantial activity and high selectivity toward the most acetylated products—DAG and TAG—when conducted under reflux conditions at 393 K. These values further improved with higher initial HAc/Gly molar ratios, leading to complete glycerol conversion and selectivity for DAG and TAG close to 90%.

Introducing a solid catalyst increased catalytic activity, although final selectivity levels were only slightly higher than those seen in the autocatalytic reaction. This suggests that selectivity is largely governed by thermodynamic equilibrium. Comparing solid catalysts revealed that A15 performed better than the CBV720 zeolite when added in equal mass ratios (catalyst/glycerol). However, when the amount of acid sites per gram of glycerol was considered, CBV720 matched the performance of A15, showing that the acid sites within this hierarchical zeolite were accessible to reactants and had similar strength.

Previous studies on zeolites have shown that the catalytic results for H-Y and H-ZSM-5 were only comparable to the autocatalytic reaction due to an inadequate choice of zeolite [18,27,31]. The hydrophilic nature of the catalyst surface can result in active site deactivation due to water formation during esterification, which may cause leaching of active components into the reaction medium. In the present work, the CBV720 catalyst, with its hydrophobic surface from a high Si/Al ratio, resists water-related deactivation and is effective for glycerol acetylation. In addition, their porous structure prevents catalyst deactivation shown by other solid catalysts by partial blockage of active sites by glycerol or glycerides within the pore structure of catalysts. In this paper, the selected CBV720 zeolite has a higher Si/Al ratio than previous works; it is more hydrophobic, allowing the exit of the water molecules from the channels and probably preserving the acidity of the sites. These cited differences explain the better results obtained.

Conversely, CBV300 failed to reach similar conversion rates, even with a fivefold increase in acid site concentration per gram of glycerol, likely due to restricted accessibility to its acid sites.

We would expect that diffusion problems would be worse on CBV300 due to their microporosity. On the other hand, the water formed upon esterification may be more harmful to their acid sites, affecting their activity to catalyze the reaction. As CBV300 has a lower Si/Al ratio than CBV720, it is more hydrophilic, not allowing the exit of the water molecules from the channels and probably affecting the acidity of the sites

In summary, the study shows that achieving a catalyst with overall performance comparable to the commercial resin (A15) depends on loading a comparable amount of accessible acid sites, and an adequate pore structure and hydrophobicity. The lower activity of CBV300 appears to stem from limited accessibility. Nonetheless, selectivity for DAG and TAG was above 90%.

The investigation into p-toluene sulfonic acid revealed that using homogeneous catalysts could match the activity and selectivity of heterogeneous systems (A15 or CBV720), but with twenty times fewer protons. This is attributed to the higher mobility of the free acid and the transport limitations imposed by the pore network in the solids. Despite this, solid catalysts offer important advantages in terms of environmental impact and ease of final separation. Furthermore, CBV720 catalyst can be reused, maintaining their activity and selectivity across at least five cycles. There are no studies that analyze the reusability of H-Y zeolites in glycerol acetylation.

Regarding the kinetics, the autocatalytic reaction requires 24 h to reach a yield of 75% of DAG+TAG. Under equal amount of acid sites, that level is reached at 18.8 h with CBV300, while CBV720 and A15 require approximately 12.5 and 10 h, respectively. With PTSA, this time is reduced to only 30 min. For comparison, mesoporous SBA-15 requires 60 to 250 min (1 to 4.2 h) to reach that reaction advance [44], with an activity closer to PTSA. Clearly, the pore network of the zeolites imposes some restrictions on the mobility of the molecules, although this effect is not as marked in faujasites as it was seen in MFI, where the activity was close to the blank test [42,43].

MFI zeolites have two different types of channels, straight-circular (5.3 × 5.6 Å) and sinusoidal-elliptical (5.1 × 5.5 Å), while FAU zeolites have supercages of 11.2 Å diameter, connected by 7.3 Å openings [48]. Therefore, the pores in the latter are big enough to allow the molecules to react, but impose diffusive restrictions that hinder the reaction rate.

On the contrary, SBA-15 has cylindrical mesopores of 90–100 Å, in an hexagonal-2D arrangement, and A15 has macropores of approximately 300 Å. Since the nature of the acid sites is similar, the comparatively inferior catalytic activity of the resin can be attributed to its much higher site density due to its lower surface area. In consequence, the proximity of the sites is conducive to steric hindrances, where an adsorbed molecule obstructs nearby sites as well, reducing the number of effectively available sites. It is worth noticing that esterification proceeds via an Eley-Rideal mechanism, where the acid is first protonated and then reacts with the alcohol group via a nucleophilic attack; in consequence, neighboring sites are not required [58].

As a conclusion, in order to reach catalytic activity with a solid acid, that could be comparable to the activity of a homogeneous acid such as PTSA, the accessibility as well as the density of the sites are to be taken into account. Although FAU improves these properties with respect to MFI and reaches an activity comparable to A15, they fall short of achieving the activity of homogeneous and some mesoporous catalysts.

Finally, the selectivity for triacetin is constrained by water produced during the three acetylation steps. Improvements could come from employing catalytic systems that continuously remove water or use an entrainer, as previous studies have shown [59,60]. Another promising approach is the use of catalytic solids that both supply acid sites and absorb the water produced. For example, expanded polyethylene functionalized with sulfonic groups has demonstrated superior results in the esterification of oleic acid with ethanol [61]. Exploring these alternative processes will be a focus for future research.