Upcycling waste cooking oil into renewable hydrocarbon fuels using CeO₂–La₂O₃–NiO mixed oxide catalysts

2.1. Materials

The study utilized several materials, including MCC (Avicel pH 101), lanthanum(III) nitrate hexahydrate (La(NO₃) ₃ ·6H₂O, 99%, Merck, Germany), cerium(III) nitrate hexahydrate (Ce(NO₃) ₃·6H₂O, 99%, Merck, Germany), and nickel(II) nitrate hexahydrate (Ni(NO₃)₂·6H₂O, 99%, Merck, Germany). Ethanol (C₂H₅OH, 96%, Merck), citric acid (99%, Sigma Aldrich), and deionized water were also used. WCO collected from the ITS lecturer housing area was employed as the feedstock and consisted mainly of unsaturated fatty acids (40.02% oleic acid) and saturated fatty acids (34.72% palmitic acid).

2.2. Synthesis of CeO₂ -based oxide catalysts (CeO₂–La₂O₃–NiO)

The CeO₂–La₂O₃–NiO mixed oxide catalysts were synthesized via an in-situ method templating process. Ce(NO₃) ₃·6H₂O, La(NO₃)₃·6H₂O, Ni(NO₃)₂·6H₂O, and citric acid were dissolved in stoichiometric molar ratios (1.0; 0,5; 0,5M) using a solvent mixture of 30 mL ethanol and 10 mL deionized water, followed by magnetic stirring at ambient conditions for 12 h to ensure complete complexation of the metal ions. MCC (Avicel 101) was incorporated as the hard template by dispersing the required amount of MCC in 20 mL of distilled water and sonicating the suspension for 1 h before addition. The MCC dispersion was then introduced into the precursor solution at three loading levels, 12.5%, 25%, and 37.5% (w/w relative to total metal precursors), and the mixture was stirred at 60°C for 24 h to facilitate uniform in-situ templating. The resulting suspension was evaporated at 80°C to obtain a viscous gel, dried at 70°C for 24 h, and calcined at 300°C for 4 h in a muffle furnace to remove the cellulose template and form the mixed oxide network. The final catalysts were designated as BOCe (no MCC), BOCe-12.5%, BOCe-25%, and BOCe-37.5% in accordance with the MCC content used during synthesis.

2.3. Material characterization

A comprehensive set of characterization techniques was employed to examine the physicochemical properties of the synthesized materials. Thermogravimetric analysis (TGA, Hitachi STA7200) was used to determine the optimal calcination temperature, in which approximately 100 mg of pre-calcined sample placed in an Al₂O₃ crucible was heated from room temperature to 1000°C at 10°C min^-1 under an air atmosphere, and the mass-loss profile was recorded. X-ray diffraction (XRD, PANalytical X’PertPRO) with Cu Kα radiation (λ = 1.5406 Å, 40 kV, 30 mA) was performed over a 2θ range of 1–80° to identify the crystalline phases, supported by pattern evaluation using X’Pert HighScore software. Fourier transform infrared spectroscopy (FTIR, Shimadzu 8400S) was utilized to identify functional groups by collecting spectra in the 4000–400 cm^-1 region after preparing KBr pellets containing 1–2 wt% of finely ground sample. Transmission electron microscopy (TEM, Hitachi HR-9500) was conducted to observe pore arrangement, microstructure, and lattice characteristics; samples were ultrasonically dispersed in isopropyl alcohol and deposited onto copper grids prior to imaging at magnifications of 50,000–600,000×. Surface morphology and elemental distribution were further examined using field emission scanning electron microscopy coupled with energy dispersive x-ray spectroscopy (FESEM–EDX, Hitachi S4800) following Pd/Au coating for 15 min at a vacuum pressure of 6 × 10^-2 mbar. Textural properties were analyzed using a Micromeritics TriStar II 3020 surface area analyzer equipped with three parallel analysis ports, enabling simultaneous nitrogen adsorption–desorption measurements. Before analysis, 0.2 g of each sample was degassed at 300°C for 3 h under vacuum; the specific surface area was determined using the Brunauer–Emmett–Teller (BET) method, and pore size distribution using the Non-localized density functional theory (NLDFT) method.

2.4. Evaluation of catalytic performance in waste cooking oil deoxygenation

The deoxygenation of WCO was carried out in a semi-batch reactor equipped with a heating mantle and magnetic stirring. A catalyst loading of 1 wt% was introduced into 10 g of WCO inside a three-neck flask reactor. Prior to commencing the reaction, nitrogen gas was purged through the reactor system for 10 min to establish an inert atmosphere, followed by maintaining a continuous N₂ flow of 20 mL min⁻¹ throughout the reaction. The mixture was continuously stirred and heated to 360°C, where the reaction was maintained for 4 h. The vapor-phase products were routed to a condensation unit maintained at 18°C to ensure efficient condensation of the liquid fraction. The conversion and liquid yield are presented in Eqs (1) and (2). The condensed liquid products were analyzed using a GC–MS system (HP 6890 GC) fitted with an HP-5MS capillary column (30 m × 0.25 mm × 0.25 μm). Quantitative analysis was conducted using 1-bromohexane as the internal standard. Product yields and molecular composition were determined from the integrated chromatographic peak areas. Hydrocarbon selectivity was assessed by comparing the C₁₅–C₁₇ fractions in the WCO feed and the resulting liquid products, as calculated using Eq. (3). A schematic illustration of the deoxygenation setup is presented in Figure 1.

Close

Figure 1.

Deoxygenation reactor for biofuel production.

Export to PPT

3.1. Thermogravimetric analysis and derivative thermogravimetry (TGA-DTG)

Figure 2 presents the TGA–DTG profiles of the CeO₂–La₂O₃–NiO precursors containing different MCC loadings. All samples show an initial weight loss of approximately 5–8% within 50–200°C, corresponding to the removal of physisorbed and crystalline water [16]. The major decomposition event occurs between 200–380°C, where metal nitrates undergo exothermic degradation accompanied by the release of NOₓ and oxygen. In MCC-templated samples, this region exhibits a larger mass loss (35–50%) compared to the template-free precursor (22–25%), due to simultaneous thermal degradation of cellulose. During this stage, MCC undergoes depolymerization and pyrolytic cracking, generating volatile fragments that are subsequently oxidized to CO₂, contributing to the intensified weight-loss profile [17].

Close

Figure 2.

TGA–DTG profiles of (a) BOCe (without MCC), (b) BOCe–12.5 wt% MCC, (c) BOCe–25 wt% MCC, and (d) BOCe–37.5 wt% MCC.

Export to PPT

A further exothermic decomposition stage is observed between 380–600°C, resulting in an additional 12–35% mass loss as residual cellulose-derived char and carbonaceous species are oxidized. This process coincides with the progressive formation of mixed metal oxide phases. Above 600°C, only minor mass changes occur, indicating advanced structural stabilization and crystallization of the CeO₂–La₂O₃–NiO framework. As expected, the total mass loss increases proportionally with MCC loading, from ∼35% in the non-templated sample to nearly ∼70% in the 37.5% MCC precursor, reflecting the larger fraction of organic matter introduced by the cellulose template. This multistep decomposition pattern aligns with previous reports on cellulose-assisted oxide synthesis and citrate–nitrate combustion systems, where sequential water removal, nitrate breakdown, cellulose oxidation to CO₂, and final oxide formation are consistently observed [18].

3.2. X-ray diffraction (XRD)

Figure 3 displays the XRD patterns of the CeO₂–La₂O₃–NiO catalysts synthesized with varying MCC contents. All samples exhibit well-defined diffraction peaks characteristic of fluorite-type cubic CeO₂ (JCPDS 34-0934), including reflections at 2θ ≈ 27.8°, 32.3°, 46.6°, and 55.3°, corresponding to the (111), (200), (220), and (311) crystallographic planes, respectively [19]. The persistence of these reflections across all compositions confirms the formation of a stable cubic CeO₂ phase and indicates that MCC incorporation does not induce any phase transformation [20,14]. However, systematic broadening and attenuation of peak intensities with increasing MCC content suggest improved dispersion of CeO₂ crystallites and a larger amorphous fraction resulting from the templating effect of cellulose. The noticeable broadening observed at 37.5 wt% MCC further implies enhanced structural disorder and smaller coherent scattering domains.

Close

Figure 3.

XRD patterns of (a) BOCe, (b) BOCe-12.5% MCC, (c) BOCe-25% MCC, (d) BOCe-37.5% MCC.

Export to PPT

No distinct reflections associated with La₂O₃ or NiO were observed in any sample. This absence can be rationalized by two complementary factors. First, both La³⁺ and Ni²⁺ possess ionic radii compatible with the CeO₂ fluorite lattice, enabling partial incorporation into CeO₂ during calcination and leading to substitutional solid-solution formation rather than segregated oxide phases. Second, the nominal concentrations of La and Ni are below the typical XRD detection threshold (<3–5 wt%), and any weak reflections are likely masked by the dominant CeO₂ peaks or fall within the background noise level [21-23]. Such behavior is consistent with prior reports on rare-earth- and transition-metal-modified ceria, where dopant phases remain XRD-silent at low loadings due to lattice incorporation and peak overlap. These results collectively indicate successful integration of Ni and La into the CeO₂ matrix while preserving the overall cubic phase structure.

3.3. Fourier transform infrared (FTIR)

The FTIR spectra of the calcined CeO₂–La₂O₃–NiO catalysts have been shown in Figure 4. A broad band centered at ∼3445 cm⁻¹ corresponds to O–H stretching vibrations, indicating the presence of surface hydroxyl groups. These hydroxyl species are commonly retained on rare-earth oxides even after high-temperature calcination and can significantly influence catalytic performance by modulating acid–base characteristics and facilitating adsorption or activation of oxygenated intermediates during deoxygenation reactions. Bands observed at ∼1484 and 1382 cm⁻¹ are assigned to surface carbonate (CO₃²⁻) species, originating from spontaneous CO₂ chemisorption on basic oxide sites. The formation of carbonates is typical for Ce- and La-containing oxides due to their high affinity for CO₂, and these species can act as indicators of surface basicity, which plays a role in promoting decarboxylation pathways [15].

Close

Figure 4.

FTIR spectra of (a) BOCe, (b) BOCe-12.5% MCC, (c) BOCe-25% MCC, (d) BOCe-37.5% MCC catalyst after calcination.

Export to PPT

Additional features at 1064 and 853 cm⁻¹ correspond to C–O stretching and bending vibrations, respectively, arising from residual carbonate species or trace organic fragments remaining after the decomposition of MCC during calcination. The coexistence of hydroxyl and carbonate groups suggests a mixed acid–base surface environment, which is beneficial for deoxygenation reactions, as basic sites facilitate C–O bond cleavage via decarboxylation, while surface hydroxyls can assist in intermediate stabilization and hydrogen transfer processes [24].

3.4. Transmission electron microscopy

Figure 5 presents the TEM images of the BOCe catalyst, showing nanoscale particles with a tendency to form clusters. The observed agglomeration is attributed primarily to the high surface energy of the CeO₂–based nanoparticles and possible sintering during the calcination step, which is a common behavior in rare-earth oxide systems rather than being caused by residual citric acid, as most organic components are fully decomposed above 350–400°C during thermal treatment. At higher resolution (scale bar: 5 nm), clear lattice fringes can be identified, with an interplanar spacing of 0.197 nm, corresponding to the (111) crystallographic plane of fluorite-structured CeO₂ [25]. This value is consistent with previously reported d-spacings for cubic CeO₂ nanoparticles. The presence of well-defined lattice fringes confirms the high crystallinity of the CeO₂ domain within the mixed-oxide network.

Close

Figure 5.

TEM images of BOCe-25 wt% MCC catalyst: (a) agglomerated mixed-oxide nanoparticles; (b) HRTEM image showing lattice fringes (d = 0.197 nm) indexed to the (111) plane of cubic CeO₂; (c) SAED pattern indicating the crystalline fluorite CeO₂ structure.

Export to PPT

3.5. Field emission scanning electron microscopy (FE-SEM)

FE-SEM characterization was employed to investigate the structural features and surface morphology of the synthesized materials. Observations conducted at a magnification scale of up to 1 μm revealed sponge-like particles with distinct pore structures in all sample structures [17]. As shown in Figure 6(a), the BOCe sample without MCC exhibited large, irregular pores, likely caused by the breakdown of citric acid into CO₂ during the calcination process [26]. On the other hand, Figure 6(b) illustrates that the addition of 12.5 wt% MCC resulted in the formation of smaller and more numerous pores, although the pore distribution remained uneven when compared to the MCC-free sample. As shown in Figure 6(c), the BOCe-25 wt% MCC catalyst exhibits a relatively compact and densely packed morphology, consisting of closely interconnected mesoporous domains with partially constricted pores. Further increasing the MCC content to 37.5 wt% Figure 6(d) results in a highly porous and more uniform sponge-like structure with a high density of small, interconnected pores, highlighting the role of MCC as a natural hard template in promoting structured pore formation. These morphological features indicate that higher MCC loading enhances pore distribution and textural uniformity rather than pore size, which is in good agreement with the N₂ adsorption–desorption isotherm analysis.

Close

Figure 6.

FESEM images of catalysts: (a) BOCe, (b) BOCe-12,5wt% MCC, (c) BOCe-25wt% MCC, and d) BOCe-37,5wt% MCC.

Export to PPT

The EDX results, presented in Figure 7, confirm that all BOCe samples, regardless of MCC incorporation, contain Ce, La, and Ni metal elements uniformly distributed throughout the material.

Close

Figure 7.

EDX analysis of catalysts a) BOCe, b) BOCe-12,5wt% MCC, c) BOCe-25wt% MCC, and d) BOCe-37,5wt% MCC

Export to PPT

3.6. Nitrogen adsorption-desorption analysis

The N₂ adsorption-desorption isotherms of CeO₂-based oxide catalysts (CeO₂-La₂O₃-NiO) synthesized with different proportions of MCC are illustrated in Figure 8(a), with data on specific surface areas and pore size distributions provided in Table 1. According to the IUPAC classification, the isotherms for all catalysts fall under type IV, characterized by H3 hysteresis loops, which confirm the materials’ mesoporous nature. The presence of H3 hysteresis loops suggests the existence of slit-like pores, typically associated with non-uniform or irregular pore structures, likely arising from MCC serving as a hard template during the synthesis process [24].

Close

Figure 8.

(a) Isotherm graph and (b) pore distribution of CeO₂-based oxide materials (CeO₂-La₂O₃-NiO).

Export to PPT

At relative pressures (P/P₀) below 0.31, nitrogen initiates pore filling by forming a monolayer within the material. As the relative pressure surpasses 0.41, hysteresis loops emerge, indicating capillary condensation and the gradual adsorption of nitrogen into the mesopores. Table 1 presents the specific surface area data calculated using the BET method. The results reveal that the BOCe-25 wt% MCC sample possesses the largest surface area (52.36 m² g^-1), followed by BOCe-37.5 wt% MCC (48.25 m² g^-1), BOCe-12.5 wt% MCC (36.00 m² g^-1), and pure BOCe (33.41 m² g^-1). This trend demonstrates that MCC addition significantly enhances surface area, with the optimal concentration being 25 wt% MCC.

Conversely, mesopore volume data (Table 1) indicate a gradual decline in pore volume as MCC concentration increases. The pure BOCe sample features the highest pore volume (0.137 cm³/g), whereas values decrease progressively to 0.077 cm³ g^-1 in BOCe-12.5 wt% MCC, 0.079 cm³ g^-1 in BOCe-25 wt% MCC, and 0.072 cm³ g^-1 in BOCe-37.5 wt% MCC. This reduction suggests that MCC serves as an effective template, facilitating the formation of a highly structured mesoporous network. During calcination, MCC is gradually removed, leaving behind newly developed pores that align with the MCC imprint, thereby yielding a more refined and uniform pore arrangement. Figure 8(b) presents the pore size distribution of materials incorporating different concentrations of MCC. In BOCe without MCC, the average pore size is 13.68 nm, which notably declines to 6.72 nm in BOCe-12.5 wt% MCC, further decreasing to 5.27 nm in BOCe-25 wt% MCC, before exhibiting a slight increase to 5.39 nm in BOCe-37.5 wt% MCC. This pattern suggests that utilizing MCC as a hard template effectively reduces pore size while improving the uniformity of porosity, with the optimal structure achieved at 25% MCC content.

Overall, MCC functions as a natural templating agent, significantly influencing the mesoporous characteristics of the synthesized material. Its presence facilitates the formation of a structure with smaller, more controlled pores and enhanced uniformity in pore distribution, making it a valuable component in tailoring the physicochemical properties of CeO₂-based oxides.

3.7. Deoxygenation reaction of waste cooking oil using CeO₂-based oxide catalysts

The deoxygenation reaction of WCO was carried out using CeO₂-based oxide catalysts with different MCC concentrations. A semi-batch reactor setup was employed, equipped with a magnetic stirrer, a digital heating mantle, a condenser, a thermometer, and a vacuum system. The reaction involved mixing 10 g of WCO with 0.1 g of catalyst in a three-neck round-bottom flask. This process was maintained at 360°C for 4 h under a nitrogen atmosphere to prevent oxidation. The resulting liquid product was analyzed using Gas Chromatography-Mass Spectrometry (GC-MS).

As illustrated in Figure 9(a), all catalysts demonstrated exceptional efficiency, achieving 100% conversion of WCO, [27], previously reported a comparable conversion rate for algae oil when processed with lanthanum-based catalysts. The liquid product yield ranged from 57% to 33%, with the following performance ranking: BOCe > BOCe-12.5 wt% MCC > BOCe-25 wt% MCC > BOCe-37.5 wt% MCC. The highest yield of 57% was obtained using BOCe, likely due to its larger mesoporous volume [13]. The mesopore-to-micropore volume ratio in BOCe was significantly higher than in other catalysts, improving reactant accessibility and enhancing product diffusion, thus optimizing mass transfer during deoxygenation.

Close

Figure 9.

(a) Liquid yield, (b) Product selectivity, (c) Carbon selectivity, and (d) Hydrocarbon selectivity of various catalysts.

Export to PPT

Figure 9(b) depicts the formation of oxygenated intermediates, such as carboxylic acids, alcohols, and ketones, alongside non-oxygenated compounds, including hydrocarbons, cyclic structures, and aromatic species. Hydrocarbon selectivity was notably high across all catalysts, with BOCe-12.5 wt% MCC yielding the highest hydrocarbon content at 99%, followed by catalysts with MCC concentrations of 0 wt% (98%), 37.5 wt% (94%), and 25 wt% (90%). These findings reinforce the catalytic efficiency of these materials in facilitating deoxygenation. Similar results were reported by Khalit [28], who documented an 89% hydrocarbon yield using nickel-based catalysts on selected supports. Previous research on bimetallic Ni-Ce catalysts has identified cerium oxide (CeO₂) as a crucial component in enhancing catalytic activity. Furthermore, the elevated hydrocarbon yield observed with BOCe can be linked to its larger mesopore diameter (13.68 nm), which is greater than that of other samples with smaller pore sizes. A wider mesopore diameter promotes efficient diffusion of bulky WCO molecules into the active sites of the catalyst, strengthening interactions between the catalyst’s acidic sites and WCO, thereby optimizing hydrocarbon selectivity [29]. However, increasing MCC concentration in this study led to a reduction in pore size, restricting reactant accessibility by limiting molecular diffusion to the catalyst surface. Consequently, the structural and dimensional characteristics of catalyst pores play a vital role in determining reactant diffusion efficiency, decomposition kinetics, and product formation within the catalytic matrix, directly influencing the overall performance of the deoxygenation reaction [30,31].

Analysis of Figure 9(c) reveals that the highest concentration of hydrocarbon fractions falls within the C15–C17 range. This trend is primarily driven by decarboxylation and decarbonylation reactions, which eliminate carbon atoms by releasing CO₂ or CO molecules [32]. The presence of C15 and C17 fractions corresponds to the free fatty acid (FFA) components found in WCO, predominantly oleic acid (C18:1) and palmitic acid (C16:0), both of which undergo decarboxylation and decarbonylation (DCO₂/DCO) transformations [13,27]. As illustrated in Figure 9(d), paraffin formation exceeds olefin production, indicating that decarboxylation is the dominant reaction mechanism [33]. The selection between decarboxylation and decarbonylation is influenced by the degree of WCO accessibility to catalytically active sites, which subsequently impacts hydrocarbon formation. In the BOCe catalyst, which features the largest pore diameter (13.68 nm), WCO molecules diffuse more efficiently into active regions, fostering robust acidic interactions that favor alkane hydrocarbon synthesis through decarboxylation. Conversely, the BOCe-37.5 wt% MCC catalyst, characterized by a smaller pore size (5.39 nm, Table 1), primarily generates alkene hydrocarbons via decarbonylation, as restricted diffusion limits reactant access to catalytic sites.

Additionally, an increased paraffin-to-olefin ratio is closely linked to hydrogenation, which occurs as part of the overall reaction mechanism, further shaping hydrocarbon distribution and selectivity [16].