Obtention and Products Distribution of Bioliquid from Catalytic Pyrolysis of Tomato Plant Waste

Abstract

The use of tomato plant residues (i.e., stems, leaves, etc.) as a substrate for catalytic pyrolysis of biomass was investigated. A comprehensive study was conducted to investigate the impact of catalysts on the performance of different pyrolysis fractions (i.e., gas, biosolid, waxes, and bioliquid) as well as the distribution of products within the bioliquid. The catalysts employed in this study were derived from two distinct types of zirconia. The first type was synthesized by a conventional sol-gel method, while the second type was prepared with a modified method aimed at improving the presence of mesopores. This modification involved the incorporation of Pluronic 123. These materials were designated ZrO₂ and ZrO₂P25, respectively. Both types of zirconia were used as supports for tungstophosphoric acid (H₃PW₁₂O₄₀, TPA), a heteropolyacid with a Keggin structure, in the preparation of catalysts with strong acid sites. The results demonstrated that the bioliquid yield of the non-catalytic fast pyrolysis of tomato plant waste was approximately 23% and that the obtained bioliquid contained a wide variety of molecules, which were detected and quantified by GC-MS. In the presence of the catalysts, both the bioliquid yield and the distribution of bioliquid products were substantially modified. Furthermore, the possible sugar degradation pathways leading to the formation of the molecules present in the pyrolytic bioliquids were thoroughly examined. The results obtained from this study indicate that the physicochemical characteristics of the catalysts, specifically their pore size and acidity, have a significant impact on the selectivity of the catalytic processes towards valuable molecules, including anhydro-sugars and furanic derivatives such as furfural and furfuryl alcohol.

1. Introduction

Waste lignocellulosic biomass is mainly composed of hemicellulose, cellulose, lignin, and volatile components. Hemicellulose is an amorphous polymeric fraction consisting mainly of C5 and C6 sugar units. Cellulose is a biopolymer consisting of glucose (C6) units, while lignin is an interconnected structure of phenolic units. The treatment of lignocellulosic waste by chemical and/or catalytic processes is of great interest in light of the principles of the circular economy, which proposes the reuse of waste to obtain products that can be reintroduced into the production chain (or to generate products of industrial interest). These products could be used as a source of energy or as a source of platform molecules. Biomass waste can be transformed in several ways, e.g., through biochemical or thermochemical processes. Biochemical processes, such as digestion or fermentation, are those that are assisted by microorganisms or enzymes. Thermochemical processes, on the other hand, involve chemical transformations at high temperatures and in controlled atmospheres. Gasification, combustion, liquefaction, carbonization, and pyrolysis are all thermochemical processes [1].

Pyrolysis is a process in which thermochemical transformations take place in a controlled atmosphere, which may be inert or reactive (assisted pyrolysis). Pyrolytic processes are divided into three categories: slow pyrolysis, fast pyrolysis, and flash pyrolysis. This classification is based on some operational factors, mainly the heating rate and the residence time [2]. Classical fast pyrolysis is defined as the thermal decomposition of the substrate in the presence of an inert gas, such as nitrogen or argon, under oxygen-free conditions and at intermediate heating rates (seconds) [3]. The products of fast pyrolysis of lignocellulosic biomass are a gaseous fraction called biogas, a solid fraction called biochar, a waxy fraction, and a liquid fraction called bioliquid or bio-oil. The latter two fractions have a high water content (around 15–30%) and contain organic molecules, mainly oxygenated compounds. In this particular context, it should also be noted that there are numerous reports of pyrolysis systems with flow reactors being employed whose operating conditions do not fully correspond to any of the classical classifications [4]. For instance, there are so-called fast pyrolysis works, where the substrates are subjected to high heating rates but are subsequently maintained at the desired temperature for periods ranging from 5 to 10 min, with the objective of enhancing the yield of the process [5,6,7,8,9]. This particular method of thermal processing was employed in the experimental design of the present study.

The composition of bioliquids is extremely complex. They can contain more than 100 molecules from different organic families, and molecules from these families can contain C, H, O, N, and S in their structure. Short-chain aldehydes and ketones, furanic compounds, anhydro-sugars and their derivatives, carboxylic acids and alcohols, and lignin-derived aromatic and phenolic compounds are among the most common oxygenated compounds in bioliquids. Moreover, most of these molecules are reactive and can combine with each other. Pyrolytic liquids can be used directly as an energy source in boilers or subjected to upgrading processes such as hydrogenation, cracking, and polymerization to produce fuels and/or valuable chemicals [1,2,3].

While pyrolytic liquids can be used in a wide variety of industries, including cement, steel, distillery, and power generation, applications are still very limited due to the low stability and high acidity of these highly reactive bioliquids [10,11]. To overcome these limitations, heterogeneous catalysts can be used during the pyrolysis process. Catalytic assistance can modify the yield towards a fraction and/or selectively direct the pyrolysis towards a family of compounds. Changing the selectivity towards a product or a family of products is linked to controlling the physicochemical properties of the catalysts, e.g., their acidity.

Catalyst-assisted pyrolysis can be carried out by two different methods. The first method is known as ‘in-situ’ catalytic pyrolysis and consists of mixing the catalyst with the biomass. Conversely, in the ‘ex-situ’ catalytic pyrolysis process, the primary vapors emanating during the pyrolysis process are subsequently brought into contact with the catalyst, which is kept in a bed separate from the biomass (downstream). This latter method is particularly advantageous in terms of the ease of separation of the catalyst and the solid fraction (biochar or biosolid).

Among other catalysts, zirconium oxide (ZrO₂) exhibits catalytic properties suitable for use in pyrolysis reactions. Properties that make it suitable include its high thermal stability, its surface properties associated with its polymorphism, and its controllable acid-base and textural properties [12]. These characteristics of ZrO₂ allow the design of a catalyst capable of promoting depolymerization, cracking, and/or dehydration reactions. Thus, for instance, zirconia was used by Behrens et al. to promote catalytic pyrolysis towards furanic compounds [9]. Moreover, the acid-base properties of ZrO₂ can be modified by the addition of cations or anions to its structure. For example, the addition of sulfate or tungstate ions leads to highly acidic materials [13,14]. Alternatively, zirconia has also been used as a support for structures that confer acidity to the catalyst, such as Keggin heteropolyacids. These acids have been used to increase the acidity of polymeric matrices such as ZrO₂, SiO₂, and TiO₂ [15,16,17,18].

Keggin-type heteropolyacids (HPAs) are protonic acids containing anions with a complex structure known as heteropolyanions. These compounds exhibit significant Brönsted-type acidity and thermal stability, making them suitable for applications requiring high acidity and elevated temperatures. They also exhibit considerable solubility in polar solvents and a low specific surface area (~5 m²/g). These characteristics are considered limitations that hinder their effective utilization. These disadvantages can be avoided by anchoring or immobilizing heteropolyacids on a solid support with a high specific surface area and suitable pore size.

In this context, zirconia is a favorable option for use as a catalytic support for heteropolyacids, mainly due to the fact that zirconia can be prepared by a variety of methods that address its textural properties (micro-, meso-, and macroporosity). For instance, it may be obtained by sol-gel, micellar, or mechanochemical synthesis [19]. Among these methodologies, the preparation of ZrO₂ by the sol-gel method, using zirconium alkoxide as a precursor [15], is particularly promising. A key advantage of this method is its adaptability, which allows modification by incorporation of ionic or neutral surfactants. In this regard, Pluronic P123 (poly(ethylene glycol)-block-poly(propylene glycol)-block-poly (ethylene glycol)) has been used as a mesopore-forming agent in zirconia [20]. Furthermore, it is possible to engineer these materials with a bimodal pore distribution, incorporating micro- and meso-pores, a design that has the potential to facilitate the diffusion of reagents and promote catalytic activity.

In this work, zirconia and heteropolyacid-modified zirconia catalysts are studied to assist biomass pyrolysis. The biomass selected for this study comprises seasonal residues from tomato plants, including leaves and stems. The selection of this substrate is based on the scarcity of literature reports related to the obtention of bioliquids from the treatment of these wastes. Moreover, the low number of studies based on the use of tomato plant residues is particularly noteworthy considering the global nature of tomato production. In 2022, world tomato production reached 186.2 million tons, with a total cultivated area of 5 million hectares [21]. It has been estimated that between 15,000 and 30,000 tomato plants can be planted per hectare. When the dry matter generated per plant is taken into account, it can be predicted that between two and five tons of dry stubble can be generated per hectare [22]. It is estimated that global tomato production generates between 10 and 25 million tons of stubble (lignocellulosic residues) per year [23]. According to a 2023 report by the National Institute of Agricultural Technology (INTA) of Argentina [24], based on data provided by FAOSTAT in 2021, the main tomato-producing countries were China, India, and Turkey, which produced more than 10 million tons of tomatoes per year and cultivated ~2.2 million hectares. The report further indicates that in Argentina, the average annual production of tomatoes has been approximately 1 million tons and 17,000 hectares in recent years. Of this total, 3000 hectares are located in the province of Buenos Aires. The area around the capital of this province, La Plata and its surroundings, is the most important horticultural production area in the country, with around 1500 hectares of different varieties of tomatoes (round, plum, and cherry).

As is evident, the amount of biomass that is available for potential use is very large. The use of biomass wastes is of very high importance as a secondary economic revenue pathway for agriculture, which is an economic area of traditionally low added value. Incineration and the use of raw residues or ashes for soil improvement are among the practices traditionally used to treat waste [25]. Although incineration can be employed as a means of generating energy and ash that can be used for soil enrichment when conducted in a suitable manner, the open burning of waste is more common in rural areas and is typically associated with the release of pollutants into the atmosphere. In recent years, there have been reports of potential applications of biomass involving technology-based developments, such as bipolarization [26,27] and biogas production through anaerobic processes [28]. In this sense, pyrolysis can be regarded as an alternative and complementary practice, given that it requires minimal sophisticated equipment and can be readily scaled up and applied in rural environments. The process of pyrolysis can be conducted in a shorter time compared to anaerobic treatments and could also provide three valuable fractions (biogas, bioliquid, and biochar) that can be used for energy production [29], to obtain high value-added molecules, or to obtain biochar that could be used for soil improvement [30] or as an adsorbent for pollutants present in aqueous sources [31,32]. In summary, the process of pyrolysis has the potential to yield economic resources with the capacity to influence the local economy. Moreover, it can offer the benefits associated with the disposal and reuse of waste in an environmentally sustainable process. Consequently, the management of waste from this agricultural activity has the potential to generate resources and contribute to the resolution of environmental issues.

Among the few works that have studied the pyrolysis of tomato waste, a report by Font et al. stands out [33]. This report presents a kinetic study of the thermal decomposition of agro-industrial tomato waste at temperatures below 450 °C. The experiments were conducted using a combination of thermogravimetric and mass spectroscopy-coupled-thermogravimetric techniques. The authors analyzed the evolution of some reaction products in the gas phase. Prasat and Murugabelh’s analysis of the thermochemical cracking of tomato skins revealed the presence of highly flammable compounds within the pyrolytic liquids [34]. In a more recent work, Li et al. studied the pyrolysis of tomato leaves to obtain biochar [35]. Elkhalifa et al. reported the potentiality of obtaining pyrogases containing appreciable amounts of H₂ and CH₄ from the pyrolysis of vegetable waste, such as tomato, cucumber, carrot, and their blend [29]. The same authors also reported the production of biochar from the remains of the same vegetables [36]. Ozbay et al. conducted a study on slow and fast catalytic pyrolysis of industrial tomato processing residues using alumina-supported oxidized copper catalysts. These researchers established that the catalysts slightly enhanced the yield towards the liquid fraction compared to non-catalytic pyrolysis [37]. Furthermore, they observed a slight modification in the composition of the bioliquid. In another study conducted by the same research group, Fe/Al₂O₃ catalysts were employed for the same process, and it was determined that these catalysts favor the yield towards the gaseous fraction [38].

In this context, the present paper details the synthesis and characterization of zirconia and tungstophosphoric acid-modified zirconia and their use in the catalytic fast pyrolysis of tomato plant residues (stems and leaves). The main objective of this research is to obtain bioliquids from pyrolysis tests and to analze the effect of the presence of different catalysts on the yield of the liquid fraction and on the composition of the bioliquids. To achieve this objective, two ZrO₂ samples were prepared: one by the sol-gel method and another by the same method but modified with Pluronic 123 to obtain a catalyst with a higher proportion of mesopores. Subsequently, both zirconia samples were impregnated with tungstophosphoric acid (TPA). The composition of the obtained liquids was analyzed by gas chromatography coupled to mass spectrometry (GC-MS). Furthermore, from the molecules detected in the bioliquids, possible fragmentation routes obtained in the presence and absence of a catalyst were proposed.

2. Results

2.1. Catalyst Characterization

Both XRD patterns of ZrO₂ produced using a sol-gel approach (ZrO₂) or a modified method employing Pluronic 123 as a mesopore former (designated ZrO₂P25) exhibited a broad peak at 2θ~32°, indicative of low zirconia crystallinity (see Figure 1). Conversely, the XRD pattern of the ZrO₂P25TPA sample exhibited a series of broad and weak peaks between 2θ = 25–30°. No characteristic peaks of H₃PW₁₂O₄₀-21H₂O (TPA-21H₂O) or those corresponding to hydrated TPA, H₃PW₁₂O₄₀-6H₂O (TPA-6H₂O), were observed. Given the temperature employed during the heat treatment phase of the catalyst preparation procedure, the observed peaks may be ascribed to the presence of small H₃PW₁₂O₄₀ crystals. However, the intensity of the weak peaks suggests that TPA is mainly either dispersed on the support as an amorphous phase or as a crystal with a size that cannot be detected by this technique [16].

Textural properties were analyzed by N₂ adsorption/desorption, and the results achieved are summarized in

Table 1. Textural properties of the synthesized catalyst.

Samples	SBET (m2/g)	SMIC (m2/g)	SMESO (m2/g)	Dp (nm)	Vp * (cm3/g)
ZrO2	213	162	51	5.4	0.136
ZrO2TPA	87	25	62	6.9	0.094
ZrO2P25 *	192	15	177	3.6	0.178
ZrO2P25TPA *	90	6	84	5.0	0.088

* Estimated from the value corresponding to P/P₀ = 0.98.

.

The N₂ adsorption-desorption isotherms at −196 °C of ZrO₂ and ZrO₂ TPA (Figure S1, Supplementary Material) are typical of micro-mesoporous materials and can be classified as type IV with an H2 hysteresis loop, according to IUPAC classification [39]. The ZrO₂ sample exhibited a high specific surface area and a high proportion of microporosity (S_MIC/(S_MIC + S_MESO) = 0.76). Furthermore, the pore size distribution of this sample indicates the presence of a bimodal mesopore structure, with a maximum at 3.8 and 6.2 nm, resulting from the incorporation of TPA into ZrO₂, leading to a reduction in the specific surface area. This fact can be associated with a marked drop of microporosity (S_MIC/(S_MIC + S_MESO) = 0.29). The decrease in the microporosity could be attributed to the micropore blockage by the TPA anion, whose size is 1.2 nm [40].

Likewise, the isotherms of ZrO₂P25 and ZrO₂P25TPA samples are typical of mesoporous material (Figure S1c,d). The specific surface area of ZrO₂P25 is comparable to that of the ZrO₂ sample. As a result of the Pluronic 123 incorporation during the synthesis, the specific surface area ascribed to the present mesopores increases and accounts for more than 90% of the total surface. Additionally, the P123 incorporation generates an increase in the pore volume and modifies pore size distribution (Figure S1c). The ZrO₂P25 sample presents mainly small mesopores (with a maximum at ~2.8 nm). The surface area of the ZrO₂P25TPA sample is reduced by 50% compared to the bare sample. This decrease can be mainly attributed to the partial blockage of the mesopores.

The FT-IR spectrum of the ZrO₂ sample (Figure 2) exhibited an intense band between 1595 and 1343 cm⁻¹, assigned to the stretching vibrations of hydroxo and aquo-OH and to the bending vibration of the Zr-O-H species present in the solid structure, respectively [41,42,43]. In addition, in the energy range below 850 cm⁻¹, the broad band associated with the Zr-O stretching vibration is observed. A significant decrease in the intensity of the bands assigned to the -OH groups was observed in the FT-IR spectra of ZrO₂TPA and ZrO₂P25TPA. The spectrum of the ZrO₂P25 sample displays the aforementioned bands in addition to three weak signals at 1081–1055 and 850 cm⁻¹ (which are attributable to C-O-C/C-O-H stretching and CH₂ rocking vibrations, respectively) and which are characteristic of the P123 template [44]. This finding indicates that the ultrasound treatment did not completely remove P123.

As previously reported, the FT-IR spectrum of TPA displays bands located at 1081, 982, 888, 793, 595, and 524 cm⁻¹, which are assigned to the P-Oa, W-Od, W-Ob-W, W-Oc-W stretching vibrations, and Oa-P-Oa bending vibration, respectively [45,46]. Subscripts indicate oxygen atoms bridging between W and P heteroatoms (a), oxygen atoms sharing vertices (b) and edges (c) belonging to WO₆ octahedra, and terminal oxygen (d) in the TPA structure. Whereas the FT-IR spectrum of bulk Na₇PW₁₁O₃₉ shows bands at 1100, 1046, 958, 904, 812, and 742 cm⁻¹ and a shoulder at 861 cm⁻¹ [47]. The intense band at 1080 cm⁻¹ in the spectrum of the [PW₁₂O₄₀]³⁻ anion is assigned to the ν3 vibration of the central PO₄ tetrahedron. In the spectrum of the [PW₁₁O₃₉]⁷⁻ anion, this band is split into two components (1100 and 1046 cm⁻¹) due to a decrease in the symmetry of the PO₄ tetrahedron. In the case of the ZrO₂TPA sample, the FT-IR spectrum reveals the presence of the P-Oa, W-Od, W-Ob-W, and W-Oc-W TPA bands, which overlap those of the support [48]. As illustrated in Figure 2, the FT-IR spectrum of ZrO₂P25TPA exhibits the characteristic bands of the lacunar [PW₁₁O₃₉]⁷⁻ species at 1101, 1059, 957, and 818 cm⁻¹ in addition to the first three bands [47]. The [PW₁₂O₄₀]³⁻ anion could undergo partial transformation into these species during the synthesis and drying steps due to the limited stability of the Keggin anion [20]. At pH 1.5–2, the [PW₁₂O₄₀]³⁻ anion reversibly and rapidly converts into the lacunar [PW₁₁O₃₉]⁷⁻ species as the hydroxyl concentration increases according to the following equilibrium [49]:

[PW₁₂O₄₀]³⁻ ⇔ [P₂W₂₁O₇₁]⁶⁻ ⇔ [PW₁₁O₃₉]⁷⁻

In summary, the results of the FT-IR analysis have confirmed the successful incorporation of the [PW₁₂O₄₀]³⁻ anion into both the ZrO₂TPA and ZrO₂P25TPA materials. In the case of ZrO₂P25TPA, it was partially transformed into the lacunar anion [PW₁₁O₃₉]⁷⁻.

The acidity of the synthesized materials was estimated from acid-base potentiometric titration using n-butylamine as a titrant (Figure 3). The strength of the acid sites was classified according to the following scale: Ei > 100 mV (very strong sites), 0 < Ei < 100 mV (strong sites), −100 < Ei < 0 (weak sites), and Ei < −100 mV (very weak sites) [50,51]. Additionally, Ei (the initial electrode potential) indicates the maximum acid strength. The number of acid sites was calculated by measuring the area under the curve. The results obtained are summarized in Table S1 of the Supplementary Material.

In summary, the results of FT-IR analysis have confirmed the successful incorporation of the [PW₁₂O₄₀]³⁻ anion into the ZrO₂TPA and ZrO₂P25TPA materials. In the case of ZrO₂P25TPA, it was partially transformed into the lacunar anion [PW₁₁O₃₉]⁷⁻.

The ZrO₂TPA catalyst (Ei = 683 mV) exhibited higher acid strength compared to the other catalysts. The acidity of the ZrO₂ support (Ei = 62 mV) was found to be remarkably lower, and the presence of TPA in the impregnated sample was evident since the bulk TPA has an Ei of 690 mV. The ZrO₂P25TPA catalyst presented acid sites (Ei = 41 mV), but its acidic strength was found to be considerably lower compared to that of the ZrO₂TPA catalyst. This observation may be attributed to the low acidic strength of the ZrO₂P25 support, which exhibited an Ei of −30 mV. The low acidic strength of this support is ascribed to the base treatment (NaOH) used during the preparation method to remove P123. It should be noted that the presence of sodium ions on this support was corroborated by X-ray photoelectron spectroscopy (XPS) (see Figure S2 of the Supplementary Material). The interaction between TPA and the zirconia matrix is assumed to be electrostatic in nature, due to the occurrence of proton transfer to the Zr-OH groups, as evidenced by the following equation:

Zr-OH + H₃PW₁₂O₄₀ → [Zr-OH₂⁺]_n[H_3-nPW₁₂O₄₀]ⁿ⁻

However, hydrogen bond interactions between the oxygen atoms of TPA and the hydroxyl groups of zirconium could also occur. As a consequence of heat treatment at temperatures above 300 °C, Zr-O-W bonds can be formed between TPA and ZrO₂ [52] through:

[Zr-OH₂⁺]_n[H_3-nPW₁₂O₄₀]ⁿ⁻ → [Zr-O-W-PW₁₁O₃₉]²⁻ + H₂O

The lower acid strength observed in the samples containing supported TPA, compared to that of bulk TPA, could be ascribed to the occurrence of protons in TPA as H+(H₂O)₂ species. In contrast, within the supported samples, protons interact with oxygen in the Zr-OH or are present as water molecules, a consequence of the heat treatment. It is evident from the FT-IR spectra of the ZrO₂TPA and ZrO₂P25TPA samples that the decrease in the intensity of the bands assigned to -OH species is indicative of the occurrence of -OH groups as water molecules.

Regarding the number and type of acid sites determined for the catalysts, it was observed that the ZrO₂TPA catalyst exhibited the highest number of strong and very strong acid sites, which were associated with the presence of the heteropolyacid with its structure capable of donating protons in an acid-base equilibrium, as clearly evidenced in the titration curve (Figure 3).

Therefore, it can be concluded that the addition of TPA to the ZrO₂ and ZrO₂P25 supports modifying their acidic properties. The enhancement of both the acid strength and the total number of acid sites was found to be more pronounced when ZrO₂ was utilized as the starting support (refer to Table S1 in the Supplementary Material).

2.2. Biomass Characterization

2.2.1. Proximate Analysis, Moisture, and Ash Content

The residual biomass used as substrate in the present study (stems and leaves of tomato plant residues) contained 27.5% cellulose, 16.5% hemicellulose, and 8.2% lignin. This composition was determined by the Van Soest method. The moisture and ash content were determined to be 10.2% and 10.4%, respectively. The remaining percentage is made up of the extractive components such as proteins, flavonoids, fatty acids, terpenes, and others, which can be degraded during pyrolysis.

2.2.2. Preliminary Results by TGA

The biomass sample was analyzed by thermogravimetry (TGA) to determine the mass loss as a function of temperature. The thermogram obtained and the curve of the derivative of the thermogram are shown in Figure 4.

The thermograms of the several biomass samples present a first stage of mass loss between RT and 100–115 °C, associated with the loss of adsorbed water. As can be seen in Figure 4, the thermogram of the tomato biomass sample shows this mass loss (~10%). It should be noted that the water mass loss observed in this analysis is comparable to that determined through moisture analysis techniques. Subsequently, within the temperature range of 115 to 190 °C, a minimal mass loss of approximately 3% was recorded. This is due to the degradation of volatiles and other low molecular weight degradables. At higher temperatures, a significant asymmetric and unresolved signal was observed between 190 and 430 °C, corresponding to a mass loss of 38%. A comprehensive analysis of the derivative of the TGA curve reveals that this signal is composed of at least two contributions: hemicellulose and cellulose degradation, whose highest derivative rate occurs at 321 °C. The hemicellulose decomposes more easily than cellulose, and, according to literature, this process can generate a high concentration of volatile components, including acetic acid, CO, CO₂, and light hydrocarbons, among others. This mass loss signal was coupled with the mass loss signal associated with cellulose degradation. Cellulose is more stable and has a degrading temperature range of 200–450 °C, with its maximum degrading rate typically observed between 290 and 380 °C [53]. At temperatures in excess of 450 °C, mass loss processes become observable, corresponding to the decomposition of the most stable lignin fraction and the concomitant formation of ash. In this sense, the lignin degrades slowly over a wide temperature range (180–900 °C), and a part of this chain, known as refractory lignin, has been shown to degrade at high temperatures (>450 °C) [54]. As shown in the thermogram in Figure 4, the temperature ranges in which the different fractions decompose overlap, making it difficult to quantitatively associate the mass loss of each step with the decomposition of a particular polymer [53,55,56,57]. In addition, from the results obtained, it can be concluded that a temperature of 450 °C would be suitable for pyrolysis tests, as at this temperature the sugar chains are considered to be depolymerized/degraded. The percentage mass loss as determined in this study up to 450 °C is closely aligned with the percentage contributions of hemicellulose and cellulose as determined by proximate analysis. Given the objective of this research is to obtain pyrolytic derivatives from the sugar chains, thermograms were not conducted until temperatures exceeded 700 °C.

2.2.3. Efficiency

Figure 5 shows the percentage of each fraction (gas, bioliquid, biosolid, waxes) obtained in the pyrolysis tests both in the presence and absence of catalysts. The waxy fraction consists of those compounds capable of volatilizing and solidifying in the condenser. This fraction contains, in addition to small molecules (e.g., phenolic compounds, 1-hydroxy-2-propanone, ethylhydroxyacetate, methylbenzene, 2-cyclopenten-1-one, 2-furanmethanol, etc.), condensation products, high molecular weight molecules, and/or fatty acids, among others. These solid/waxy phases can be diluted with acetone, and the presence of the aforementioned compounds was revealed by gas chromatographic analysis coupled with mass spectrometry. However, as the analysis of the waxy fraction was not part of the objective of this work, the composition of this fraction was not studied in depth. The solid fraction, known as tar, contains biochar and dark condensates and is made up of a mixture of polycyclic compounds, large molecules, and other components [58]. Finally, the gaseous fraction, calculated by difference, consists of non-condensable gases. The present study has focused exclusively on a detailed analysis of the composition of the liquid fraction, which is referred to as “bioliquid”.

The efficiency of each fraction in the pyrolysis test performed in the absence of a catalyst is comparable to that observed in the presence of bulk tungstophosphoric acid (TPA). Only a slight increase in the percentage of the gaseous fraction was observed, which is likely due to the presence of the acid sites of TPA [59]. On the other hand, the ZrO₂-assisted test showed an increase in bioliquid production compared to the non-catalytic test. This suggests that the catalyst may be involved in the transformation of condensable gases into the compounds that form the liquid fraction. The observed increase in the liquid fraction was accompanied by a decrease in wax generation. A substantial enhancement in the yield of the solid fraction, tar, was also observed. This tar is produced through a series of polymerization and/or self-condensation reactions [60,61].

In the experiment conducted in the presence of the ZrO₂TPA catalyst, the most acidic catalyst used in this study, a decrease in the production of bioliquid was observed, along with an increase in the yield of the gaseous fraction. As previously reported by Zou et al. and Inayat et al., these gases arise from the fragmentation of molecules formed during pyrolysis, catalyzed by strong acid sites [62,63]. It is worth mentioning that the impact of this catalyst on the efficiency of the different fractions was more significant than that observed with bulk TPA. This is likely due to the fact that the ZrO₂TPA catalyst displays a higher specific surface area and contains the TPA dispersed on the catalytic surface, as evidenced in the characterization studies.

Conversely, the test carried out with the ZrO₂P25 catalyst showed a slight increase in yield towards the condensable wax phase. Furthermore, a substantial decrease in the yield of the solid fraction was observed. These results suggest that certain compounds that typically remain in the solid fraction, known as tar, may undergo a cracking process, resulting in the formation of smaller compounds, which subsequently solidify in the condenser. This catalyst has the largest surface area, characterized by the presence of mesopores, which may facilitate access of larger molecules to the active sites. Additionally, the ZrO₂P25TPA catalyst, which has low porosity and some acid sites on its surface, promoted the formation of solids, thereby increasing the yield of that fraction [64].

2.2.4. Bioliquid Composition

Biomass pyrolysis can be described in several stages. First, heat transfer occurs from the heat source to the biomass. Next, there is an increase in temperature, leading to the release of volatile compounds that will come into contact with the non-pyrolyzed particles, increasing the heat transfer. The substrates to be pyrolyzed are degraded in this medium, generating the so-called primary pyrolysis products. Part of these products may react, leading to side reactions, which contribute to the further conversion or decomposition of such compounds [65,66].

In general, pyrolysis reactions are known to be very complex. They can be classified as primary or secondary reactions, with primary reactions involving the main pathways of depolymerization, fragmentation, dehydration, and carbon formation. Subsequently, in secondary reactions, the volatile compounds originating from the primary reactions undergo cracking into smaller compounds or recombination into compounds with higher molar mass [67]. The correct choice of pyrolysis temperature is essential to determine the quantity and quality of molecules that can be obtained. In the present study, the pyrolysis temperature was selected based on the findings of the preliminary thermogravimetric test. At the selected temperature (450 °C), the polymeric sugar chains (hemicellulose and cellulose) were completely degraded, while lignin was barely degraded.

Figure 6 presents the composition of the bioliquids derived from non-catalytic and catalytic pyrolysis of tomato plant waste (leaves and stems) in the form of relative content (percentage) per group. The compounds determined for each group, along with their relative abundance, will be presented subsequently in this paper. An initial analysis revealed that all pyrolytic tests, irrespective of whether they were conducted with or without a catalyst, resulted in the detection of significant quantities of acids, ketones, and furanic compounds in the bioliquids. The majority of these compounds were identified as degradation products of sugars. A significant proportion of the bioliquid products resulting from the non-catalytic experiment have been found to be comprised of acids and ketone compounds, accounting for more than 62% of the sample. However, it should be noted that the presence of the catalysts has the effect of modifying the relative percentages found for each group. These results are consistent with those reported by other researchers when using different lignocellulosic substrates (corn cob, grasses) [68,69].

A thorough analysis of the compounds present in the bioliquids was conducted to assess the impact of the catalysts on the number of carbons of the compounds obtained. Figure 7 illustrates the number of carbons of the compounds obtained. These results are derived from the analysis of all the compounds present in the bioliquid, where the number of molecules containing a specific number of carbon atoms is determined. As seen in this figure, most of the products obtained have a carbon number of six or less. Non-catalytic pyrolysis preferentially generated C2 compounds; however, this distribution varied substantially in the presence of catalysts. In the case of the ZrO₂ catalyst, the formation of C5 and C6 compounds was favored, among which some are phenolic in nature, as previously mentioned. The use of the ZrO₂P25 catalyst resulted in the preferential formation of larger molecules, particularly C6 compounds and some larger ones (C12). These results agree with the observations made during the analysis of the efficiency of each pyrolysis fraction obtained, suggesting that this material, which contains larger pores and low acidity, allowed the formation of larger compounds, some of which remain in the bioliquid and others in the form of waxes.

The bioliquids obtained in the pyrolytic tests performed in the presence of TPA did not show significant differences with respect to the bioliquid obtained in the non-catalytic test. The former presented a slightly higher content of C4 and C5 compounds, mainly furanics, aldehydes, alcohols, and esters, as previously analyzed.

The ZrO₂TPA catalyst, which exhibited the highest acidic strength and the highest number of strong acid sites, significantly promoted the formation of C5 compounds. As will be demonstrated in a subsequent section of this study, the primary constituents are furanic compounds. Finally, the ZrO₂P25TPA catalyst does not appear to direct the pyrolysis process towards a specific carbon number.

A closer analysis of the results reveals that the presence of catalysts seems to favor different degradation routes. While the non-catalytic experiment revealed a predominance of ketones and acids, this behavior was not replicated for all catalysts. To illustrate this point, if the role of the ZrO₂P25TPA catalyst is considered, it can be observed that, in the presence of this catalyst, anhydrosugars constitute the predominant family of compounds within the bioliquid. This finding can be attributed to the mesopore size of the catalyst, since mesoporosity has been shown to provide a favorable environment for the entry of levoglucosan (LGA), promoting its dehydration [70]. On the other hand, experiments conducted in the presence of ZrO₂ demonstrated an enhancement in the generation of lignin derivatives, accompanied by a large increase in aldehyde content and a decrease in acid concentration. These compounds are derived from the cleavage of sugar polymers at specific acidic sites, which are likely to be insufficient in ZrO₂. In contrast, in the presence of the ZrO₂TPA catalyst, the production of furanic compounds is favored. These compounds are generated by acid site-assisted dehydration reactions. Furthermore, it was noted that the incorporation of acid sites via TPA impregnation resulted in a concomitant increase in the fragmentation of the compounds and the production of acids.

In general, and as expected, high concentrations of lignin derivatives were not observed. The presence of phenolic and benzene compounds within the bioliquid composition can be attributed to the decomposition of aromatic compounds, which are constituents of the non-refractory lignin fraction of the biomass [71]. This assumption is based on the finding of syringol, guaiacol, and eugenol in the chromatographic analysis of the bioliquids. These molecules are the primary derivatives of lignin decomposition. However, the individual concentration of these compounds never exceeded 4% of the bioliquid composition. This is likely due to the low temperature used during pyrolysis and the low concentration of lignin found in the tomato leaves and stems (8.4%).

In order to facilitate the analysis of the results found in the pyrolysis tests, a scheme (Scheme 1) is shown in this paper that outlines the most widely accepted and potential degradation pathways for polymeric sugar chains, monomers, and their derivatives. This scheme has been formulated based on empirical results reported in the literature and theoretical studies using DFT (Density Functional Theory) [72,73,74,75,76,77,78,79]. Pyrolysis predominantly facilitates dehydration reactions, thus promoting the cleavage of glycosidic bonds between the polysaccharides present in the biomass. As described by Shen et al., the thermal decomposition of cellulose is described as follows. The 1,4-glycosidic bond in the cellulose polymer is cleaved, forming anhydrosugars. The glucose monomers then undergo a rearrangement process, resulting in the production of lactic acid, acetic acid, hydroxyacetone, and furfural, among other compounds [78]. It is important to mention that Scheme 1 also proposes some pathways (via a–d) to obtain compounds that can be derived from side reactions. These reaction pathways were incorporated following the detection of multiple compounds in the bioliquids through chromatographic analysis, which were not exclusively produced by cleavage reactions and/or dehydrations. For instance, the process leading to the reduction of furfural to its more reduced forms, such as furfuryl alcohol and tetrahydromethylfurfural (via c, Scheme 1), or to oxygenated compounds, e.g., 2(5H)furanone [80].

In addition, some of these pathways can be promoted by the presence of catalysts. In this work, the catalyst was placed in a catalytic bed separated from the biomass bed. In this arrangement, the catalysts may cause some of the primary favored, accelerating the transformation of intermediates into full degradation products, such as acetic acid, acetol, 2,3-butanedione pathway, etc. Conversely, it is also possible that some pathways can be disfavored, in which case the obtaining of an intermediate product, for example, levoglucosanone (via c, Scheme 1), can be facilitated. Further reactions may be promoted, leading to cracking, decarbonylation, carbonylation, esterification, condensation, and/or polymerization of the compounds present in the pyrolysis vapors, which contain numerous highly reactive compounds at high temperature. The progress of the different catalytic reactions depends on the physicochemical characteristics of the catalyst, such as porosity, surface area, acid-base properties, etc. [81].

In the following, a comprehensive analysis of the compounds found in the bioliquids in each of the pyrolytic tests (with and without catalyst assistance) is carried out. To organize the analysis, the contributions of the particular constituents of each family are identified. Figure 8 displays the different anhydrosugars that were produced during the pyrolysis tests, along with their respective percentage concentrations in the bioliquids.

Despite the low concentration of compounds belonging to this family in the bioliquids, their detection has facilitated the identification of reaction pathways in the catalytic reactions. It is also worth noting that most of the anhydrosugars detected in the bioliquids from the catalytic pyrolysis test were not detected in the bioliquids from the non-catalytic test. For the production of anhydrosugars, the catalyst that presented a distinctive result was ZrO₂P25. It is evident that this catalyst significantly hindered the degradation of intermediate products, resulting in the generation of substantial concentrations of anhydro-sugars. ZrO₂P25 showed a remarkable increase in the production of anhydrosugars (23%) compared to the result found in the non-catalytic test. This finding could be attributed to the contribution of the catalyst mesopores, which enabled the depolymerization of holocellulose (i.e., cellulose and hemicellulose). This process resulted in the generation of primary sugars via the cleavage of 1,4-glycosidic bonds, subsequently leading to the formation of 1,6-anhydro-sugars [70]. As Jiang et al. suggested, cellulose appears to be more susceptible to dehydration than to hydrolysis during the pyrolysis process [82]. The same authors proposed that the dominant mechanisms for the formation of levoglucosan (LG) and its anhydrides include homolytic (radical), heterolytic (ionic), and/or concerted mechanisms. The presence of LG derivatives in bioliquids is significant because they demonstrate possible reaction pathways [83]. Mainly the following anhydrosugars were obtained: levoglucosenone (LGA), 4-O-β-D-galactopyranosyl-α-D-glucopyranose, and 1,4:3,6-dianhydro-α-D-glucopyranose (via b, Scheme 1). It is worth mentioning that obtaining DAGP is of high interest due to its high cost and limited supply in the market. The importance of obtaining it is based on the fact that this compound is an intermediate for the production of medicinal drugs and compounds of interest to the food industry [84,85]. Conversely, the low acidity of the material was likely insufficient to further promote dehydration reactions and subsequent degradation to smaller molecules.

Subsequently, secondary reactions may occur, including dehydration, isomerization, fragmentation, and decarboxylation. These reactions have been shown to generate furanic compounds from cleavages and rearrangements of anhydrosugars [78], as depicted in Scheme 1, or from the cleavage of the C-O bond in pentoses, such as pyranose rings, followed by dehydration and cyclization reactions [86]. In addition to the aforementioned thermal decomposition reactions, other reactions occur in the catalyst-assisted pyrolysis process, such as hydrodeoxygenation, hydrocracking, hydrogenolysis, decarboxylation, decarbonylation, ketonization, aldol condensation, and dehydrogenation [87].

As illustrated in Figure 9, the pyrolysis tests yielded a variety of furanic compounds.

In the absence of a catalyst, the most abundant furanic compound was furfural, a known product of the dehydration of pentoses or anhydro sugar derivatives [88,89,90,91], followed by 2(5H)-furanone, which is an oxidation product of furfural. Lu and Wu analyzed pyrolytic vapors from different biomasses and determined the species present in the gas phase. Utilizing mass spectrometry, they determined that key ion fragments, such as the C₄H₅O₂⁺ ion, can be obtained from the cellulose (C6) and/or hemicellulose (C5 and C6) fractions responsible for the formation of 2(5H)-furanone and other furan derivatives [92].

Catalytic pyrolysis in the presence of bulk TPA yielded results analogous to those obtained in the absence of catalyst. This catalyst was found to slightly promote the formation of two representative compounds of the furan family, namely furfural and 2(5H)-furanone. Despite the presence of acid sites within this heteropolyacid that could potentially facilitate the formation of furanic compounds, the low activity observed can be ascribed to its comparatively low specific surface area. This, in turn, results in a reduced probability of contact between the catalyst and the gaseous molecules being produced.

When the reaction was carried out with the assistance of ZrO₂ and ZrO₂TPA catalysts, the main products were furfural and 2(5H)-furanone. In this respect, while 5% furfural was obtained in the absence of the catalyst, more than double the yield (approximately 12%) was obtained in the presence of the ZrO₂TPA catalyst. As already mentioned, the last catalyst was the one with the highest acid strength and the highest number of total acid sites. These acid sites were found to promote the dehydration reactions of the sugar monomers and to favor the cleavage and isomerization reactions of the various anhydrosugars formed first [86,93]. Furthermore, with both catalysts, furfuryl alcohol was observed, which is a product of the selective reduction of furfural (via c, Scheme 1). With the ZrO₂TPA catalyst, the formation of furfuryl alcohol was much more pronounced. It should be noted that no furfural reduction products were observed in the non-catalytic test. John B. Paine et al. conducted an investigation into the mechanisms of pyrolytic formation of furan derivatives, utilizing 13C isotopic labelling. The authors proposed that various furans arise from mechanisms involving reduction (hydrogenation). Furthermore, they postulated that the reduction process must necessarily originate from an intermolecular mechanism [94]. In this sense, furfuryl alcohol could have been obtained from a catalytic transfer hydrogenation (CTH) with an indirect source of H. The CTH could be carried out by the presence of alkyl alcohols, such as methanol, as H donor species and a furfural carbonyl group (acceptor). The reaction could be assisted by the acid sites of a catalytic surface. In addition, the low H₂ availability allows a better selectivity of the hydrogenation/hydrogenolysis reaction, thus generating the fully hydrogenated C5 product, tetrahydrofurfuryl alcohol [95]. In the context of complex mixtures such as pyrolytic bioliquids, the presence of molecules capable of participating in catalytic transfer hydrogenation (CTH) reactions, such as alcohols, cannot be overlooked. In recent years, a significant number of studies have focused on obtaining furfuryl alcohol from furfural by zirconia-assisted CTH reaction. Xu et al. reported good catalytic performance of a Zr-doped mesoporous carbon structure (OMC) catalyst [96]. Concurrent studies by other researchers have yielded analogous results, utilizing zirconia-SBA15 and defect-rich amphoteric ZrO₂ catalysts, which are characterized by a high density of surface acid-base sites [97,98,99]. These reports support the results obtained in the catalytic pyrolysis presented in this study.

A different behavior was observed when using the ZrO₂P25 catalyst. The presence of this catalyst resulted in the predominant formation of tetrahydro-2-furfuryl alcohol, a product arising from the non-selective hydrogenation of furfural. Furthermore, no evidence of furfural was detected in the pyrolytic liquid. The presence of furfuryl alcohol and 2,2,4,4-tetramethyltetrahydrofuran was also observed, although in a lower concentration. It is evident that the presence of this catalyst, which exhibits the highest amount of mesopores, facilitated the entry of furfural molecules and H₂ donor compounds into the mesopores, thereby promoting the CTH reactions. In addition, the bioliquid obtained in this pyrolytic test contained other compounds, including 5,5-dimethyl-2(5H)-furanone, which may have arisen from the oxidation of furfural (via c, Scheme 1). However, with the addition of TPA and the occlusion of larger pores, these reactions were less favored. That is the case for the ZrO₂P25TPA catalyst.

An analogous study was conducted on the composition of bioliquids in relation to the so-called carboxylic acid family. These carboxylic acids are produced by the primary decomposition of cellulose and hemicellulose, also referred to as depolymerization, which involves the cleavage of the terminal groups of the polymeric chains of sugars. The aforementioned cleavages have been observed to yield a range of products, including levoglucosan, xylan/xylose, and acetic acid, among others [68]. Alternatively, short-chain carboxylic acids, such as acetic acid or formic acid, can be produced from the degradation of anhydrosugars, with the consequent formation of reaction intermediates (acetoxyacetone/dihydroxyacetone), as presented in via a, Scheme 1. Figure S3 (Supplementary Material) illustrates the relative content of carboxylic acids in the obtained bioliquids. The non-catalytic experiment yielded approximately 23% acetic acid, with a minor contribution of formic and picolinic acids. The presence of carboxylic acids in pyrolytic liquids has been frequently observed and has been identified as the primary cause of their high acidity [11].

The bulk TPA test yielded only acetic acid, with a slightly lower content than the non-catalytic test. Analogous results were obtained employing both ZrO₂ and ZrO₂P25, yielding approximately 10% acetic acid. This suggests that these catalysts do not facilitate the reaction pathway leading to substantial intermediate fragmentation (via a, Scheme 1), as previously evidenced in this study, or it could also be that these catalysts promote reactions in which acetic acid is a reactant. Moreover, zirconia can also be said to exhibit selectivity for alternative pathways. Conversely, the utilization of catalysts comprising supported TPA, ZrO₂ TPA, and ZrO₂P25TPA yielded 16% and 22% acid, respectively. These percentages were comparable to those obtained in the non-catalytic pyrolysis. This observation indicates that the presence of TPA, in conjunction with the generation of new acid sites, can facilitate consecutive cleavages [8].

The subsequent stage of the investigation consisted of the analysis of the ketones present in the pyrolytic liquids. Figure 10 illustrates only those ketones whose percentage was higher than 2%. In the non-catalytic test, the bioliquid obtained exhibited 17% hydroxyacetone, and, to a lesser extent, the formation of 2,3-butanedione, 2,3-pentanedione, and 1,2-cyclopentanedione was detected. These compounds are derived from the intermediate acetoxyacetone (Scheme 1), which is the degradation product of 1,4:3,6-dianhydro-a-D-glucopyranose (DAGP) [79,100,101].

The catalysts examined in this study exhibited a concordant tendency to form ketones and acids, which was an expected result, as evidenced by via a, Scheme 1. Within the ketone family, the most abundant product was hydroxyacetone (acetol), which is formed through degradation reactions of anhydrosugars (via a and via b, Scheme 1) [88,102]. In particular, the bioliquid obtained in the bulk-assisted pyrolysis of TPA exhibited a reduced percentage of ketone compounds, a finding that is consistent with the high percentage of acetic acid present in this bioliquid. In a similar manner, in the presence of ZrO₂ and ZrO₂P25 catalysts, ketone formation was not favored. This finding is consistent with previously presented results for these catalysts, which indicated that they promoted alternative reaction pathways, such as the furfural formation (via b, Scheme 1). On the other hand, a notable production of dihydroxyacetone was detected in the bioliquid derived from the pyrolysis assisted by the ZrO₂P25TPA catalyst, accompanied by a remarkable production of acetic acid, as previously outlined.

It is worth noting that no results were observed in this work indicating that the reaction would have been directed towards ketonization. This reaction is known to promote the formation of acetone from acetic acid. Furthermore, analysis of bioliquids obtained in the presence of catalysts, particularly those obtained with ZrO₂ and ZrO₂P25, reveals a lower acetic acid concentration when compared to the bioliquid obtained from non-catalytic pyrolysis. This decrease is accompanied by a decline in the hydroxy-ketone concentration, suggesting that alternative reaction pathways have been promoted, most likely via pathways b or c as illustrated in Scheme 1.

Concerning aldehyde formation, it was observed that only TPA and ZrO₂ catalysts significantly promoted this process, with the formation of pentanal, and to a lesser extent, the ZrO₂P25 catalyst, with the formation of heptanal. The findings of this study are presented in Figure S4 (see Supplementary Material).

Due to their importance as building blocks, it is essential to study the impact of catalyst utilization on alcohol production in the context of pyrolysis-derived bioliquids. The various alcohols obtained in the pyrolytic tests carried out in the present study, along with their respective concentrations, are illustrated in Figure 11. In the non-catalytic test, methanol and maltol were obtained at low concentrations. In the presence of catalysts, a notable enhancement in efficiency towards alcohols was observed. For instance, the use of the ZrO₂P25TPA catalyst resulted in the detection of a methanol concentration higher than 5%. Conversely, the pyrolytic test assisted with the ZrO₂P25 catalyst yielded a bioliquid comprising longer-chain alcohols, such as pentanol and nonanol. As illustrated in Figure 10, this catalyst produced the highest percentage of furfuryl alcohol. These findings are of considerable significance, given the high potential of these alcohols for use as fuels and/or fuel additives.

Regarding esters, no molecules incorporating this functional group were identified in the pyrolytic liquid from the non-catalytic test. However, these molecules were detected in the pyrolysis liquids obtained with the TPA-containing catalysts, although their contribution was not significant (see Figure S5 in the Supplementary Material). The potential origins of these compounds may be the degradation of hemicellulose via lactic acid, as illustrated in Scheme 1 via d, or alternatively, from side reactions.

Finally, as expected (see Figure S6, Supplementary Material), the pyrolytic liquids contain low concentrations of compounds derived from this fraction. This is due to the choice of temperature for the pyrolysis tests and the low lignin content of the starting biomass. In this regard, the catalyst that exhibited the highest fragmentation capacity was ZrO₂. It was observed that the bioliquid obtained during the test assisted by this catalyst contained guaiacol, syringol, catechol, and other compounds derived from lignin [71].

2.2.5. Correlation Analysis

Correlation analysis is a statistical tool that determines the degree of relationship between two variables. In this study, Pearson’s linear correlation was used to explore the potential relationship between the major chemical compounds identified in the bioliquid resulting from pyrolysis. Particularly, this analysis is another tool that served to corroborate the suggested reaction pathways for obtaining the majority components. The correlation coefficient (r), which ranges from −1 (strong negative correlation) to 1 (strong positive correlation), was visualized using a correlogram (Figure 12), where the size and color of the circles encode the strength and direction of the relationship [103]. A correlation is considered to be strongly negative when r ≤ −0.75 with a significance level p < 0.05. This is represented by a large red circle, indicating that when the amount of compound X increases, the amount of the negatively correlated compound, Y, decreases. An example of this is shown in Figure 12, where the dashed yellow lines highlight the negative relationship between hydroxyacetone and 1,4:3,6-dianhydro-α-D-glucopyranose (DAGP). This result suggests that the formation of hydroxyacetone may be related to the degradation of DAGP, as proposed in Scheme 1. On the other hand, a strong positive correlation is considered when r ≥ 0.75 with p < 0.05. These cases are represented by blue circles in Figure 12 and indicate that two compounds have a direct relationship: when the amount of compound X increases, the amount of compound Y increases. An example of this correlation can be observed between acetic acid and hydroxyacetone. The model predicts that these two compounds may share a pathway in this case.

Among the statistically accepted correlations, the following stand out as positive: hydroxyacetone and acetic acid (r = 0.87, p = 0.0256), 2(5H)-furanone and furfural (r = 0.95, p = 0.0040), 1,2-cyclopentanedione (1,2-CD) and pentanal (r = 0.81, p = 0.0485), tetrahydrofurfuryl alcohol (THFA), and 1,4:3,6-dianhydro-a-D-glucopyranose (DAGP) (r = 0.98, p = 0.0009). On the other hand, the negative correlations found are between hydroxyacetone and DAGP (r = −0.82, p = 0.0478), furfural and DAGP (r = −0.75, p = 0.0835), 2(5H)-furanone and DAGP (r = −0.84, p = 0.0354), and THFA and 2(5H)-furanone (r = −0.81, p = 0.0492).

The findings of the statistical study are consistent with the proposals related to the reaction pathways followed during pyrolysis. The model proposes that DAGP plays a significant role in the formation of hydroxyacetone, furfural, and 2(5H)-furanone, as indicated in routes a and c of Scheme 1 and in the subsequent discussion of results. Additionally, the presence of THFA is associated with DAGP. Furthermore, the model suggests that there is a relationship between furfural and 2(5H)-furanone, with a strong positive correlation indicating a possible equilibrium reaction between these compounds. Analogous relationships were observed between hydroxyacetone and acetic acid and between 1,2-CD and pentanal. These observations suggest the presence of specific formation pathways or competitive interactions during the process of pyrolysis. Guo and co-workers reported a comparable study analyzing bioliquids obtained from diverse biomass substrates and found positive correlations between furfural and 5-methyl furfural, furfural and 2-methy propanoic acid, acetic acid and propionic acid, and catechol and 2,6-methoxyphenol, as well as negative correlations between acetic acid and 2,6-methoxyphenol, etc. [104].

3. Materials and Methods

3.1. Catalyst Preparation

In a typical synthesis, zirconium propoxide (Sigma-Aldrich^®, St. Louis, MO, USA, 26.6 g) was mixed with absolute ethanol (Merck, Darmstadt, Germany, 336.1 g) and stirred for 10 min to obtain a homogeneous solution under N₂ at room temperature. Then 0.47 mL of 0.28 M HCl aqueous solution was dropped slowly into the above mixture to catalyze the sol-gel reaction.

After 0.5 h, an ethanol solution of the triblock copolymer P123 (PEO₂₀PPO₇₀PEO₂₀, Sigma-Aldrich^®, St. Louis, MO, USA) was added under vigorous stirring to act as a mesoporous template. The amount of added solution was fixed to obtain a template concentration of 25% by weight in the final material. The reaction mixture was vigorously stirred for 0.5 h at room temperature. The stirring was stopped, and the suspension was aged for 24 h at 90 °C. The solid was isolated by filtration, washed with deionized water, and the templates were removed by ultrasonic treatment at 20 °C for 1 h. The sample was named ZrO₂P25. The same procedure was employed to prepare a sample without the addition of P123 (named ZrO₂).

The ZrO₂ and ZrO₂P25 materials were used as TPA support. The impregnation was performed by contacting, at room temperature, 0.75 g of the support with 0.25 g of tungstophosphoric acid (H₃PW₁₂O₄₀) dissolved in 3 mL of water-ethanol 50% (v/v) solution to obtain a TPA concentration of 25% by weight in the final material (named ZrO₂TPA and ZrO₂P25TPA). Finally, the solids were calcined at 450 °C for 2 h under the air atmosphere. The procedure used to determine the TPA content in the calcined materials is described in the Supplementary Material.

3.2. Catalyst Characterization

The nitrogen adsorption/desorption measurements were carried out at liquid nitrogen temperature (−196 °C) using Micromeritics ASAP 2020 equipment (Norcross, GA, USA). From the obtained data, the specific surface area (S_BET) was determined using the Brunauer–Emmett–Teller model, the micropore area (S_MIC) by the t-plot method, the mean pore diameter (Dp) by the BJH method, and the pore size distribution (PSD) by the DFT (Density Functional Theory) method.

The species present in the supports and catalysts were evaluated by FT-IR, using a Bruker IFS 66 (Billerica, MA, USA) equipment with pellets of the sample in KBr, in the 400–4000 cm⁻¹ range at room temperature.

The X-ray diffraction (XRD) patterns were recorded with Panalytical X’Pert PRO (Worcestershire, UK) equipment with a built-in recorder, using Cu Kα radiation, a nickel filter, 20 mA and 40 kV in the high voltage source, and a scanning angle between 5 and 60° 2θ at a scanning rate of 1° per min.

The acid strength and the number of acid sites were estimated from the n-butylamine potentiometric titration results obtained using a Metrohm 794 Basic Titrino apparatus (Herisau, Switzerland) with a double junction electrode.

XPS experiments were performed on a Physical Electronics spectrometer (PHI VersaProbe II Scanning XPS Microprobe, Eden Prairie, MN, USA) with scanning monochromatic X-ray Al Kα radiation as the excitation source (200 µm area analyzed, 25.0 W, 15 kV, 1486.6 eV) and a charge neutralizer. The pressure in the analysis chamber was maintained lower than 2.0 × 10⁻⁶ Pa. Spectra were charge referenced with the C 1 s of adventitious carbon at 284.8 eV. The energy scale was calibrated using Cu 2p_3/2, Ag 3d_5/2, and Au 4f_7/2 photoelectron lines at 932.7, 368.2, and 83.95 eV, respectively. The collected XPS spectra were analyzed in detail using PHI SmartSoft software 4.3.1 and processed using the MultiPak 9.6.0.15 package. The obtained spectra were fitted using Gaussian–Lorentzian curves to more accurately extract the binding energies of the different element core levels. The samples were measured as received in the laboratory, without any special pretreatment.

3.3. Biomass Characterization

3.3.1. Hemicellulose, Cellulose, and Lignin Content

Lignocellulosic biomass was sourced from tomato farmers in La Plata, Buenos Aires, Argentina. The aerial roots of tomato plants (Solanum lycopersicum L.) of the local variety called ‘platense tomato’ were used. The samples were obtained through the process of sampling and quartering 10 kg of the total sample. Only the leaves and stems were used and milled to obtain particles with a diameter ranging from 0.415 to 1 mm through sifting with a sieve. The raw tomato plant leaves and stems composition of hemicellulose, cellulose, and lignin was measured with the Van Soest method [105,106] with an ANKON 200 FIBER ANALYZER (Macedon, NY, USA), and gravimetry was used to determine moisture after drying the samples at 105 °C for 24 h. The mineral constituents were assessed by quantifying the ash content through the incineration of material aliquots in a muffle furnace at 900 °C.

3.3.2. Thermogravimetric Analysis

The thermal degradation of raw biomass was studied thermogravimetrically. The instrument is composed of a thermobalance, which allows analyzing the mass loss of the biomass sample. To prevent combustion reactions, the thermogravimetric reactor was purged with He for 20 min to obtain an oxygen-free atmosphere. The experiment was carried out in Shimadzu T50 equipment (Kyoto, Japan) from room temperature to 700 °C.

3.4. Catalytic Pyrolysis

Pyrolytic experiments were carried out on tomato waste samples at 450 °C under an N₂ flow (QN₂ = 200 mL/min) in a fixed-bed metal reactor (Scheme S1, Supplementary Material). The biomass and catalysts were loaded between quartz wool layers to retain the solid products generated while allowing the vapors to escape. The experimental conditions were selected based on preliminary TGA results. The pyrolysis reactor was heated using a type of book-furnace, a vertical electric furnace equipped with a temperature programmer-controller. Approximately 1 g of biomass and 0.5 g of catalyst, separated by quartz wool, were placed inside the tubular reactor before being rapidly introduced into the isothermal zone of the reactor, previously heated to 450 °C. The heating rate was estimated to be 225 °C s⁻¹ using this procedure. The calculated residence time used for the experiments was 0.5 s. The experiment was carried out for 5 min to ensure complete pyrolysis of tomato waste. The vapors were condensed using a water/salt/ice bath to collect the bio-liquid. After the reaction, the reactor was removed from the furnace and cooled in an inert atmosphere to prevent combustion of the solid residue. Four products were obtained after the experiment: bioliquid (collected in the condenser), wax (solids in the condenser), biochar (residual in the reactor), and biogas. The bioliquids obtained from catalytic pyrolysis of biomass samples were identified and quantitatively determined with a Shimadzu GCMS-QP2010SE chromatograph coupled with a mass spectroscopy detector, provided with a Supelco SPBTM-5 column (Millipore Sigma, Burlington, MA, USA) (30 m × 0.25 mm × 0.25 μm). The quantification was carried out considering that the peak areas are proportional to the corresponding compound concentration in the sample.

The yields of bioliquid, wax, biosolid and non-condensable gas were calculated as a percentage of the weight using the following equation.

[Yield [%wt](bioliquid) = {(weight of the bioliquid collected)/(weight of biomass feed) × 100}]

(1)

[Yield [%wt](wax) = {(weight of the wax collected)/(weight of biomass feed) × 100}]

(2)

[Yield [%wt](biosolid) = {(weight of the biosolid collected)/(weight of biomass feed) × 100}]

(3)

[Yield [%wt](non-condensable gas) = 100 − ([Yield [%wt](bioliquid)] + [Yield [%wt](biosolid)])]

(4)

3.5. Stadistical Analysis

In order to find correlations between the different components present in the bioliquids, particularly those coming from the sugar chains, a simple statistical study was carried out. A multivariate study was conducted using the R-Studio software 2024.09.1 Build 394, employing the corrplot package for data visualization. The present study is based on the application of parametric correlations (Pearson correlations) to measure the linear dependence between two variables (in this work the variables are the relative contents of some selected compounds). The Pearson correlation matrix determines significant correlations between the main products in the bioliquid composition, and in this matrix the correlations are shown as a function of the Pearson correlation coefficient, p, which indicates the distance between the data points and the correlation line. For this work, correlations with p < 0.05 were considered valid.

4. Conclusions

This work detailed the synthesis and characterization of tungstophosphoric acid (TPA)-impregnated zirconia-based catalysts and their influence on the pyrolysis of tomato leaf and stem waste. Both the yield towards the liquid fraction and the composition of the obtained bioliquids were influenced by the use of ex situ catalysts in the pyrolysis experiments. In addition, the use of catalysts led to the prioritization of certain degradation pathways. In this context, non-catalytic pyrolysis preferentially produced C2 compounds, mainly acetic acid and hydroxyacetone. In the presence of the catalysts, an increase in the concentration of C5 and C6 molecules is observed. The most abundant C5 molecules are furfural, furfuryl alcohol, THFA, 1-acetoxyacetone, 2,3-butanedione, pentanal, and 1,2-cyclopentanedione, while the most abundant C6 molecule is DAGP, which is valuable for biorefining. The presence of ZrO₂ and ZrO₂TPA, on the other hand, favors routes with a higher selectivity for furanic compounds such as furfural and furfuryl alcohol. Finally, in the presence of the ZrO₂P25 catalyst, larger molecules such as anhydrosugars were obtained. A significant increase in the production of anhydrosugars such as levoglucosenone, 4-O-β-D-galactopyranosyl-α-D-glucopyranose, and 1,4:3,6-dianhydro-α-D-glucopyranose was observed compared to other components. The physicochemical properties, mainly pore size and acidity of the catalysts, were responsible for the observed selectivity towards different routes. Catalysts with highly acidic sites favored dehydration reactions.

Compounds such as furfural are recognized as high-value biomass-derived chemicals that can be converted into furfuryl alcohol and other chemical compounds with diverse industrial applications, including solvents, pharmaceuticals, resins, plastics, and perfumes. 2,3-BD is also of particular interest due to its application as a food flavoring, having been approved by the Food and Drug Administration (FDA). Its potential as a food preservative is currently under investigation, as studies have demonstrated its efficacy in preventing fungal growth and retarding the maturation process of various vegetables. DAGP has also emerged as a compound of research interest, currently produced only by pyrolytic gas isolation. Its polyether cage structure, which exhibits notable water solubility, renders it a compound of considerable interest, with potential applications in medicinal chemistry and in the food industry.

Correlation analysis results suggested that the presence of DAGP in the bioliquid is key to the subsequent formation of hydroxyacetone, furfural, 2(5H)-furanone, and THFA.