Upgrading of Rice Straw Bio-Oil Using 1-Butanol over ZrO₂-Fe₃O₄ Bimetallic Nanocatalyst Supported on Activated Rice Straw Biochar to Butyl Esters

Abstract

Bio-oil produced via fast pyrolysis, irrespective of the biomass source, faces several limitations, such as high water content, significant oxygenated compound concentration (35–40 wt.%), a low heating value (13–20 MJ/kg), and poor miscibility with fossil fuels. These inherent drawbacks hinder the bio-oil’s desirable properties and usability, highlighting the necessity for advanced processing techniques to overcome these challenges and improve the bio-oil’s overall quality and applicability in energy and industrial sectors. To address the limitations of bio-oil, a magnetic bimetallic oxide catalyst supported on activated rice straw biochar (ZrO₂-Fe₃O₄/AcB), which has not been previously employed for this purpose, was developed and characterized for upgrading rice straw bio-oil in supercritical butanol via esterification. Furthermore, the silica in the biochar, combined with the Lewis acid sites provided by ZrO₂ and Fe₃O₄, offers Brønsted acid sites. This synergistic combination enhances the bio-oil’s quality by facilitating esterification, deoxygenation, and mild hydrogenation, thereby reducing oxygen content and increasing carbon and hydrogen levels. The effects of variables, including time, temperature, and catalyst load, were optimized using response surface methodology (RSM). The optimal reaction conditions were determined using a three-factor, one-response, and three-level Box-Behnken design (BBD). The ANOVA results at a 95% confidence level indicate that the results are statistically significant due to a high Fisher’s test (F-value = 37.07) and a low probability (p-value = 0.001). The minimal difference between the predicted R² and adjusted R² for the ester yield (0.0092) suggests a better fit. The results confirm that the optimal reaction conditions are a catalyst concentration of 1.8 g, a reaction time of 2 h, and a reaction temperature of 300 °C. Additionally, the catalyst can be easily recycled for four reaction cycles. Moreover, the catalyst demonstrated remarkable reusability, maintaining its activity through four consecutive reaction cycles. Its magnetic properties allow for easy separation from the reaction mixture using an external magnet.

1. Introduction

The increasing global energy demand, fueled by population growth and industrialization, has led to the simultaneous depletion of limited fossil fuel resources and the pressing environmental crisis caused by greenhouse gas emissions [1,2]. The Organization of Petroleum Exporting Countries (OPEC), in its latest report, projects an increase in energy demand from 290.9 million barrels of oil equivalent per day in 2022 to 359 million barrels per day by 2045 [3]. This compels a systematic investigation into sustainable and renewable energy alternatives. Lignocellulosic biomass, particularly rice straw, stands out as a promising solution to these challenges. As a non-edible agricultural residue, rice straw’s abundance and economic viability make it a transformative biofuel feedstock [2,4]. This has become increasingly crucial given the environmental challenges associated with fossil fuels and the need for energy security [5]. The viability of rice straw as a biofuel feedstock has been accelerated by its abundance and cost-effectiveness [6]. While the efficient conversion of biomass into liquid fuels remains a complex challenge [7], the utilization of biomass such as rice straw for energy and biobased products represents a promising avenue in a carbon-constrained world [8]. Hence, their efficient deployment holds the potential for a substantial decrease in gas emissions, concurrently serving as an extra reservoir for renewable biofuels [9]. Rice straw biomass, derived from one of the globally prevalent staple foods, is increasingly attracting interest in the production of bio-oil. This is achieved through fast pyrolysis methods, encompassing both catalytic and non-catalytic approaches [10].

Fast pyrolysis stands as a thermochemical technique capable of producing a liquid fuel suitable for replacing fossil fuels in static heating or electrical generation applications [11]. This process involves the thermal degradation of biomass in the absence of air at high temperatures (400–450 °C) and pressures (0.1–0.5 MPa) to liquid bio-oil, a mixture of gases, and biochar [12]. The main problem associated with bio-oil produced through fast pyrolysis is the complicated chemical composition of the bio-oil. Typically, bio-oil is characterized by a diverse blend of organic compounds such as organic acids, alcohols, aldehydes, ketones, esters, sugars, guaiacols, syringols, furans, phenols derived from lignin, levoglucosans, and various other compounds [13,14,15]. The bio-oil extracted from rice straw encounters challenges characterized by high water content, a substantial concentration of oxygenated compounds (35–40 wt.%), a low heating value (13–20 MJ/kg), and poor miscibility with fossil fuels. These inherent limitations hinder the favorable properties and usability of the bio-oil, underscoring the need for advanced processing techniques to address these challenges and enhance the overall quality and applicability of the bio-oil in energy and industrial applications [13,14,16]. Furthermore, bio-oil exhibits a pronounced inclination to polymerize during storage and transportation, accompanied by significant amounts of carboxylic acids, leading to low pH values and heightened corrosiveness [17]. Consequently, the imperative arises for the introduction of new effective upgrading techniques capable of improving the negative properties of the bio-oil. These enhancements are crucial for optimizing the utilization of fast pyrolysis bio-oil in various applications, emphasizing the necessity for continued advancements in bio-oil processing [18].

Innovative upgrading methods are essential to unlock the full potential of bio-oil as a sustainable alternative fuel. These methods aim to address the limitations of fast pyrolysis and enhance the bio-oil quality, positioning it as a more versatile and sustainable energy source [19,20,21]. Various upgrading methods have been developed to improve the physicochemical properties of fast pyrolysis bio-oil, including hydrodeoxygenation (HDO), hydrogenation, emulsification, catalytic cracking, and esterification in supercritical fluids [22]. Esterification is a chemical process that converts carboxylic acids (such as acetic, formic, and propionic acids) present in bio-oil into esters by reacting them with alcohols in the presence of acid catalysts. Both Lewis and Brønsted acids can be used as catalysts to accelerate this process. Esterification significantly reduces the acidic value of bio-oil, making it less corrosive and more suitable for use as a fuel or chemical feedstock [23,24]. The esterification process offers a more sustainable route to produce esters that are suitable for applications with biodiesel [25]. The esterification process, particularly when employing a heterogeneous solid catalyst alongside HDO, emerges as the optimal approach for bio-oil upgrading, as it not only catalyzes esterification but also facilitates the deoxygenation process, enhancing the calorific value of the upgraded fuel [23,26]. Recent advancements in fast pyrolysis bio-oil enhancement include supercritical alcohol treatment, presenting an eco-friendly alternative to conventional methods like HDO and catalytic cracking. This is because supercritical alcohols possess unique solvent properties that can lead to higher bio-oil yields and good bio-oil properties and enhance the conversion of bio-oil into higher-quality fuels and chemicals [27,28]. In this study, we used an innovative approach by utilizing supercritical butanol for bio-oil upgrading through esterification with a heterogeneous catalyst. This approach addresses common challenges, such as hydrogen availability and coke formation encountered in conventional methods [29,30]. Unlike other upgrading techniques like HDO, this approach fulfills the demand for hydrogen through a hydrogen transfer from supercritical butanol facilitated by the catalyst [29,31]. The unique properties of supercritical butanol as a solvent create a homogeneous reaction environment by dissolving gases, such as hydrogen, after dissociation [28]. This will also facilitate the HDO reactions and prevent coke formation [32]. Butanol offers several distinct advantages over ethanol and methanol as a solvent for bio-oil upgrading. Its immiscibility with water simplifies solvent recovery, reducing waste and enhancing process sustainability. Its high molecular weight contributes to lowering waste and improving overall efficiency. Butanol’s reduced toxicity makes it safer for both operators and the environment. Additionally, its high energy content and low corrosivity make it an ideal choice for the effective transformation of bio-oil [33]. Building on this foundation, the comprehensive exploration by Jin [34] and Angelici [35] into butanol’s production and application, including its potential as a substitute for traditional fuels, adds further depth to our solvent selection.

In the exploration of catalytic bio-oil upgrading, conventional petroleum hydrotreating catalysts, particularly noble metals like Rh, Pd, and Pt, have been extensively studied but face limitations due to substantial production costs [36,37]. Transition metals, including Ni, Cu, Fe, Co, and Mo, offer cost-effective catalysts for both hydrodeoxygenation (HDO) and esterification reactions [38,39]. Numerous studies have documented the catalytic activity of transition metal oxides in heterogeneous hydrogenation reactions. While noble metal catalysts are generally more active for various reactions, metal oxides have gained widespread application in catalysis due to their superior resistance to poisoning [40,41]. Moreover, various combinations, such as bimetallic CoMo and NiMo on alumina, as well as bimetallic NiCo catalysts, have been investigated [42,43,44]. Among the transition metals discussed, zirconium (Zr) and iron (Fe) are the most cost-effective, abundant, and the least investigated metals. Utilizing cost-effective and readily available transition metals such as Zr and Fe in a catalyst formulation with the right proportions of ZrO₂ and Fe₃O₄ emerges a cost-effective and magnetically recoverable catalyst that effectively catalyzes deoxygenation and esterification reactions. Zirconium oxide, together with iron oxide, acts as a Lewis acid, catalyzing the esterification reaction. An additional benefit of incorporating iron oxide is its unique magnetic property, which imparts magnetism to the catalyst, enabling easy separation from the reaction mixture by applying an external magnetic field. This dual-functional catalyst offers a sustainable and efficient solution for improving the quality and value of bio-oil as an economically viable and magnetic approach for efficient bio-oil upgrading. The magnetic bimetallic nanocatalysts, endowed with both magnetic and catalytic properties, showcase promising potential in enhancing the sustainability of biofuel production processes [45]. The use of oxide-based catalysts is preferred over alloys because oxides often exhibit higher surface areas and small particle sizes, allowing good dispersion of metal oxides (Lewis acids) over the silica-rich biochar support, which improves the catalytic activity [46]. The choice of catalyst support is crucial in bio-oil upgrading, and various materials, such as biochar, metal oxides, and zeolites, have been explored for this purpose [47,48,49]. Cu–Ni catalysts supported on homologous biochar were successfully employed for lignin depolymerization without external hydrogen [50]. In certain cases, trimetallic catalysts such as MgNiMo supported on activated carbon have demonstrated effectiveness for bio-oil esterification under supercritical ethanol conditions [23]. It is noteworthy that most of these biochar-based catalysts were made from commercially purchased activated carbon, some of which had been previously used to upgrade bio-oil derived from fast pyrolysis in supercritical ethanol [23,51]. Acidic supports like zeolites and silica-alumina exhibit effectiveness in upgrading bio-oil even at atmospheric pressure without requiring hydrogen [42]. Recently, biochar has gained popularity as a catalyst support, mirroring the trend of zeolite and alumina because of its sustainability, ready availability, environmental friendliness, and cost-effectiveness [52,53].

This study employed activated rice straw biochar as a unique acidic support instead of zeolite support due to its distinctive chemical composition of high silica content, enabling successful loading of metal oxides ZrO₂ and Fe₃O₄ on its acid sites, a feature less pronounced in biochar from alternative biomass sources [54,55]. The choice of using ZrO₂ and Fe₃O₄ as bimetallic catalysts aligns with previous studies by Ma and Chen, which demonstrated the effectiveness of transition metals, including Ni, Co, and Cu, in improving the yield and selectivity of bio-oil during the hydroliquefaction of rice straw [56,57]. Further support comes from our most recently published work, emphasizing the effectiveness of upgrading bio-oil over a bimetallic catalyst (CuO-Fe₃O₄) supported on rice straw biochar in supercritical alcohol to value-added chemicals [24]. This study aims to synthesize and characterize a magnetic bimetallic oxide catalyst (ZrO₂-Fe₃O₄/AcB) supported on activated rice straw biochar for catalytic rice straw bio-oil upgrading in butanol. Characterization methods include XRD, FTIR, BET surface area, BJH pore size distribution, elemental analysis, TGA, and SEM. The resulting esterified upgraded bio-oil will undergo standard physical and chemical characterization using Karl Fisher, GCMS, Bomb calorimetry, elemental analysis, etc. The effect of time, temperature, and catalyst loads on the yield of the upgraded products will be also studied.

2. Results and Discussions

2.1. Catalyst Characterization

The textual characteristics of the produced AcB support, as well as the fresh and used ZrO₂-Fe₃O₄/AcB catalysts, were assessed using N₂ physisorption techniques. The impact of the metal oxide loading on the BET surface area, pore volume, and average pore diameter of AcB support is presented in

Table 1. Textural properties of the prepared catalyst and support.

Catalyst	BET Surface Area (m2/g)	Pore Volume (cc/g)	Average Pore Diameter (nm)
AcB support	129.48	0.311	9.60
Fresh ZrO2-Fe3O4/AcB	38.03	0.172	18.11
Used ZrO2-Fe3O4/AcB	34.58	0.125	19.47

. The findings indicate that metal oxide loadings significantly reduce both BET surface area and pore volume of the support. Notably, an inverse relationship was observed between the BET surface area and the average pore diameter. The significant increase in pore diameter after catalyst preparation is attributed to the deposition of metal oxides (ZrO₂ and Fe₃O₄) onto the surface and within the pores of the activated biochar (AcB) support. Filling and blockage of AcB micro/mesopores during metal oxide loading leads to an increase in the average pore diameter [58]. On the other hand, both fresh and used ZrO₂-Fe₃O₄/AcB catalysts maintained sufficient BET surface area for the reaction, with little to no change in average pore diameter, indicating sustained catalytic activity.

The N₂ adsorption/desorption isotherm of both fresh and used ZrO₂-Fe₃O₄/AcB catalysts are shown in Figure 1. The isotherm exhibits a type IV pattern with an H3 hysteresis loop, which is typical of meso- and micro-porous materials with a plate-like layered structure [59]. These findings align with previous studies on supported bimetallic nanocatalysts, highlighting the impact of metal loading on structural characteristics and catalytic activity [24]. Incorporating these N₂ physisorption results offers a comprehensive understanding of how the catalyst’s properties evolve during synthesis.

To assess the thermal stability and degradation characteristics of the catalysts, thermogravimetric analysis (TGA) was performed on both fresh and used ZrO₂-Fe₃O₄/AcB catalysts. The results of the TGA, depicted in Figure 2, demonstrated a weight loss of approximately 1.0% and 0.5% around ambient temperature (300 °C) for the fresh and used catalysts, respectively. These losses were mainly attributed to the evaporation of water. Significantly, both the fresh and used catalysts exhibited thermal stability and diminished degradation tendencies within the temperature range from 300 to 800 °C. These findings align with previous studies [60,61] emphasizing the crucial role of catalyst stability in biomass conversion processes. This thermal stability is a crucial attribute, ensuring the catalyst’s sustained activity and effectiveness during the bio-oil upgrading process. However, it is noteworthy that the fresh catalyst experienced a slight additional weight loss of 0.25% beyond 350 °C, which could be attributed to remnants of biomass material. This minor degradation is possibly caused by the decomposition of AcB compounds. Although some weight loss is observed, it remains within an acceptable range, and it is unlikely to significantly affect the overall performance of the catalyst. Hence, the TGA analysis of both fresh and used catalysts offered valuable insights into their thermal behavior, confirming their suitability for the bio-oil upgrading process.

The composition and the crystallinity of the prepared fresh and used ZrO₂-Fe₃O₄/AcB catalyst were determined using XRD, along with the average crystal size. As shown in Figure 3, the characterization spectrum of both catalysts displays narrow and sharp peaks, indicating high crystallinity. The XRD spectrum of both catalysts exhibits crystallinity, with diffraction peaks at 2θ of 30.2°, 35.6°, 43.5°, 53.9°, 57.4°, and 62.8°. These peaks correspond to Fe₃O₄ and are attributed to the cubic spinel unit cell, aligning with standard magnetite structural data (JCPDS file No. 05-0629) [62,63]. The Fe₃O₄ peaks correspond to the 220, 311, 400, 422, 511, and 440 lattice planes, confirming the presence of Fe₃O₄. Additionally, ZrO₂ exhibits a polycrystalline nature, with diffraction peaks at 2θ values of 30.5°, 33.2°, 35.5°, 49.7°, 50.5°, 60.5°, and 64.2°, corresponding to the 101, 002, 200, 200, 102, and 311 lattice planes [64]. Both fresh and recycled catalysts also exhibited a characteristic diffraction peak at around 2θ of 24.5°, which corresponds to the (002) plane of carbon [63]. Furthermore, diffraction peaks at 2θ of 41.1°, 49.7°, and 60.5° represent three of these distinct phases of ZrO₂: monoclinic, tetragonal, and cubic zirconia phases, respectively. These phase transformations from monoclinic to metastable tetragonal are attributed to the substitution of Fe ions with Zr ions [64]. The crystallinity and crystal size of fresh and used catalysts were determined using the Scherrer equation [65]. The crystallinity of both fresh and used catalysts was 77.1 and 75.3%, while the average crystal sizes were 18.4 and 23.9 nm, respectively. These results elucidate that the catalyst maintains its crystallinity after usage for four cycles. Hence, the XRD has confirmed the presence of both Zr and Fe oxide crystalline phases in both fresh and used catalysts.

Figure 4 illustrates the morphology and elemental composition (Zr, Fe, Si, O, and C) of both fresh and used catalysts, examined using SEM equipped with EDX. From the SEM images, the fresh ZrO₂-Fe₃O₄/AcB catalyst (a) reveals metallic oxide crystals deposited on its surfaces. The SEM-EDX elemental composition analysis of the fresh catalyst spectra (b) confirms the loading of bimetallic oxide. Peaks of Si and O indicate the expected abundance of silica (Si-O-Si) in rice straw biochar [52]. Mapping of the fresh catalyst (c–f) demonstrates well-dispersed metal oxides on the AcB support within the synthesized ZrO₂-Fe₃O₄/AcB catalyst. Similarly, the SEM image of the used catalyst (g) shows the dispersion of metallic oxide crystals on the surfaces, which appears more compact than those on the fresh catalyst. The SEM-EDX elemental composition analysis of the used catalyst spectra (h) confirms the persistent abundance of silica (Si-O-Si). Mapping of the used catalyst shows some empty spots in the (i, j, and l) images, indicating relatively less distribution of O, Si, and Zr. This may be due to the leaching of O, Si, and Zr elements during the regeneration process. On the other hand, the mapping image of Fe element (k) shows a more uniform structure, indicating less probable leaching of Fe.

The FTIR spectra in Figure 5, which focus on ZrO₂-Fe₃O₄/AcB bimetallic nanocatalyst, reveal two weak absorption bands at (3000–2840) cm⁻¹ in the fresh catalyst. These bands are slightly more pronounced in the used catalyst corresponding to C–H stretching vibrations. Similarly, another band, 1381 cm⁻¹, corresponds to bending vibrations. These findings suggest the presence of small O-H groups in the water adsorbed on the silica surface. A broad and highly intense band at 1060 cm⁻¹ corresponds to the asymmetric stretching vibrations of Si–O–Si bonds, confirming the high concentration of silica, as seen in the SEM elemental mapping. Additionally, the band detected at 801 cm⁻¹ is attributed to the symmetric stretching vibration of Si–O–Si in the rocking mode [66].

2.2. Physical Characterization of Raw and Esterified Upgraded Bio-Oils

In addressing the challenge of coke formation observed in previous studies during bio-oil upgrading, we employed 1-butanol as a solvent to mitigate coke on catalysts [67]. This strategic choice aimed to enhance the efficiency of bio-oil upgrading while overcoming coke-related issues. To assess the impact of catalysts on the upgrading process, a control experiment without catalysts was included, and the results are presented in

Table 2. Experimental design and the upgrading products yields.

Exp. No.	Catalyst (wt./g)	Temperature (°C)	Time (h)	Yields (wt.%)
				Oil	Aqueous	Solid	Gas
Control	0	300	2	52.04	37.36	0.69	9.91
1	1.80	250	3.00	47.09	39.74	3.21	9.96
2	1.80	350.00	1.00	27.43	35.5	2.25	34.82
3	0.60	300.00	1.00	46.02	39.42	1.12	13.44
4	1.80	300.00	2.00	44.74	36.01	2.88	16.37
5	1.80	350.00	3.00	27.49	32.28	3.1	37.13
6	3.00	250.00	2.00	49.09	38.16	3.22	9.53
7	3.00	300.00	1.00	41.61	39.89	3.25	15.25
8	3.00	300.00	3.00	40.2	40.56	3.71	15.53
9	1.80	250.00	1.00	49.57	39.11	2.51	8.81
10	3.00	350.00	2.00	23.57	36.05	3.25	37.13
11	0.60	250.00	2.00	49.05	38.91	3.22	8.82
12	0.60	350.00	2.00	27.07	33.07	3.75	36.11
13	0.60	300.00	3.00	44.39	37.98	2.74	14.89
14	1.80	300.00	2.00	44.06	37.43	2.25	16.26
15	1.80	300.00	2.00	44.84	35.8	3.11	16.25

, outlining the experimental design and upgrading product yields, including those from the control experiments.

Table 3. Elemental analysis and heating value results for raw and upgraded bio-oil.

EXP. No.	Elemental Analysis (wt.%)				HHV (MJ/kg)
	C	H	N	O
Raw bio-oil	54.10	7.73	1.85	36.32	21.34
1	63.77	11.00	0.76	24.47	32.15
2	61.26	11.11	0.64	26.99	31.21
3	64.12	11.16	0.88	23.84	31.81
4	66.61	11.16	0.94	21.29	33.10
5	62.62	10.36	0.72	26.3	31.69
6	63.3	11.05	0.61	25.04	30.93
7	64.58	10.9	0.91	23.61	31.79
8	64.65	11.27	0.82	23.26	31.80
9	61.13	11.11	0.77	26.99	29.63
10	62.23	10.00	0.49	27.28	31.41
11	61.13	10.67	0.48	27.72	29.59
12	60.1	10.95	0.71	28.24	30.55
13	64.51	10.98	0.56	23.95	31.82
14	65.39	10.44	0.64	23.53	32.70
15	65.87	11.17	0.79	22.17	32.88

provides elemental analysis and heating value results for raw and upgraded bio-oil, revealing a significant improvement in calorific value, a reduction in oxygen content, and an increase in carbon content compared with raw bio-oil. The initial oxygen content of rice straw bio-oil, which was 36.32 wt.%, consistently decreased to 21.29 wt.% after the upgrading process, while the carbon content increased to 66.6% (experiment 4). Furthermore, hydrogen contents in the upgraded bio-oils with the addition of the catalyst substantially increased. For instance, in experiment 4, under optimal conditions, the carbon content increased from 54.1% (without catalyst) to 66.6%, the hydrogen content rose from 7.73% to 11.16%, and the oxygen content decreased from 36.32% to 21.29%. This consistent reduction in oxygen content, along with the increase in carbon and hydrogen content, was observed across all 15 experiments. Additionally, there was a significant rise in the Higher Heating Value (HHV) from the initial 21.34 MJ/kg in raw bio-oil to as high as 33.1 MJ/kg. These outcomes underscore the efficacy of the upgrading process in enhancing energy content and reducing oxygen levels in biofuels, highlighting the potential for producing high-quality biofuels through esterification under catalyst-addition conditions [68].

Table 4. Physical properties for raw and upgraded bio-oil.

EXP. No.	Water Content	Viscosity at 40 °C Cst	Density at 40 °C (g/cm3)	pH	TAN (mg KOH/g)
Raw bio-oil	32.5	1.88	0.92	2.9	83.5
Control	12.04	3.98	1.00	3.01	78.5
1	11.66	7.69	0.97	3.47	69.7
2	9.39	5.85	0.93	3.94	61.5
3	10.84	7.46	0.96	3.69	65.6
4	9.59	7.69	0.96	3.64	66.5
5	-	-	-	4.34	59.8
6	10.96	8.01	0.97	3.48	70.6
7	10.23	6.52	0.95	3.91	61.9
8	10.03	6.87	0.94	3.74	64.7
9	11.53	7.30	0.97	3.30	73.4
10	-	-	-	4.02	60.2
11	11.48	8.14	0.97	3.50	69.2
12	9.95	-	-	3.92	60.9
13	9.62	7.23	0.95	3.71	65.2
14	9.97	7.25	0.95	3.57	66.2
15	10.47	7.17	0.95	3.55	66.4

Aqueous fraction was removed from samples.

details the physical properties of raw and upgraded bio-oil, along with results from the control experiments. Catalyst addition significantly enhanced all measured physical and chemical properties. Under catalyst-addition experiments, the quality of the upgraded bio-oil was significantly enhanced, with water content reduced from 32.5% to 9.59%, representing a substantial 60% reduction. The Total Acid Number (TAN) decreased from 83.5 to 66.5, and the pH increased from 2.9 to 3.64. These changes indicate a reduction in the acidity of the upgraded bio-oil, which in turn reduces corrosiveness and improves fuel stability [69]. Additionally, viscosity increased from 1.88 to 7.69 cSt at 40 °C.

2.3. Chemical Characterization of Raw and Esterified Upgraded Bio-Oils

The chemical composition of the bio-oil produced from the pyrolysis of rice straw biomass predominantly contains oxygenated compounds such as acids, aldehydes, ketones, furans, and phenols. The esterification-focused upgrading experiment significantly improved the chemical composition of the bio-oil. Specifically, it led to a remarkable increase in esters from 3.08% in raw bio-oil to 44.82% in the final upgraded product. There was also a significant decrease in the concentration of acids from 27.79% to 3.31%, strongly indicating that esterification and transesterification reactions were the primary reaction routes [68]. Additionally, ketonic compounds were significantly reduced from 32% in raw bio-oil to 8% in the optimal upgrading experiment (experiment 4). This reduction, along with the decrease in aldehydes and sugars, can be attributed to the reduction reactions facilitated by both butanol and the catalyst. The detailed reaction mechanism was thoroughly discussed in our previous study [24]. This enhancement is particularly noteworthy when comparing the bio-oil composition of the optimal upgrading experiment (experiment 4) with the original raw bio-oil (Tables S2 and S3). This consistent trend was observed throughout all 15 experiments with catalysts, as depicted in Table S1 and Figure 6 and validated by comparing GC/MS compounds of raw bio-oil in Table S2 and the corresponding chromatogram in Figure S1.

2.4. RSM Analysis of Ester Yield %

The upgrading and optimization of rice straw bio-oil into upgraded bio-oil were conducted using the Response Surface Methodology (RSM) [70]. The Box-Behnken design (BBD) was utilized to explore optimal experimental conditions for upgraded bio-oil production from rice straw bio-oil. The optimization of the upgraded bio-oils was conducted using three factors at three levels, requiring a total of 15 runs with three replications at the center points to evaluate the pure error. The factors, ranges, and levels of the investigated variables are listed in

Table 5. Experimental design levels with 3 factors.

Factors	Unit	Symbol Coded	Levels in Box-Behnken Design
			Low (−1)	Medium (0)	High (+1)
Temperature	°C	T	250	300	350
Catalyst loading	g	C	0.6	1.8	3
Time	min	M	1	2	3

. The optimal condition combination that will produce the highest ester yield was determined with the statistical model using OriginPro 2021b software (9.8.5.212 version, OriginLab Corporation, Northampton, MA 01060, USA).

The main factors that affected the bio-oil composition, as well as ester yield, were reaction temperature (T), catalyst loading (C), and reaction hold time (M). The catalyst reaction temperature levels ranged from 250 to 350 °C, the concentration levels were between 0.6 and 3.0 g, and the reaction times varied from 1 to 3 h. These ranges were chosen based on the previous literature surveys. After completing the experiments, the response variable (ester yield) was applied in a full quadratic model to correlate it with the independent variable.

Table 6. Experimental matrix with the predicted values of ester yield %.

Std Order	T	M	C	Experimental Ester Yield %	Predicted Ester Yield %	Error
1	250	3	1.8	43.41	43.2	0.21
2	350	1	1.8	40.52	40.73	−0.21
3	300	1	0.6	40.25	39.98	0.27
4	300	2	1.8	44.83	44.7	0.13
5	350	3	1.8	39.06	39.01	0.05
6	250	2	3	42.15	42.09	0.07
7	300	1	3	41.07	41.09	−0.02
8	300	3	3	40.95	41.22	−0.27
9	250	1	1.8	38.86	38.91	−0.05
10	350	2	3	37.81	37.59	0.23
11	250	2	0.6	38.59	38.82	−0.23
12	350	2	0.6	40.88	40.95	−0.07
13	300	3	0.6	42.44	42.42	0.02
14	300	2	1.8	44.59	44.7	−0.11
15	300	2	1.8	44.68	44.7	−0.02

shows the complete experimental design matrix of BBD for factorial design with predicted values of ester yields. To avoid systematic errors, the run order was randomized [71]. In the upgrading experiments, the ester yield ranged from 37.81% to 44.83%, with the highest yield obtained under reaction conditions of a catalyst concentration of 1.8 g, a reaction time of 2 h, and a reaction temperature of 300 °C. Based on the coded parameters, a full quadratic polynomial regression model with determined coefficients for statistical prediction, developed by RSM, correlates the ester yield response as a function of the factors (T, C, M). Table 6 presents the resulting design matrix and the values of the experimental and predicted responses. The predicted highest yield of ester was 44.7% with a 0.13 error. The results confirm that the reaction conditions of a catalyst concentration of 1.8 g, a reaction time of 2 h, and a reaction temperature of 300 °C are optimal. Figure 7 displays the 2D contour optimization plots of ester yield. The regression coefficients, calculated T-values, and p-values were summarized in Table S4.

An analysis of variance (ANOVA) was conducted to determine the statistical significance and fitness of the model equation, and the results are reported in

Table 7. Analysis of variance (ANOVA) of full quadratic polynomial model with statistics fit.

Source	DF	Sum of Squares	Mean Square	F Value	Prob > F	Remarks
T	1	2.80845	2.80845	37.06546	0.00173	Significant
M	1	3.3282	3.3282	43.92504	0.00118	Significant
C	1	0.00405	0.00405	0.05345	0.82633	Not significant
T*T	1	23.84275	23.84275	314.6727	1.04 × 10−5	Highly significant
M*M	1	6.27714	6.27714	82.8447	2.68 × 10−4	Highly significant
C*C	1	15.72577	15.72577	207.54619	2.91 × 10−5	Highly significant
T*M	1	9.03003	9.03003	119.17679	1.12 × 10−4	Highly significant
T*C	1	10.98922	10.98922	145.03398	6.97 × 10−5	Highly significant
M*C	1	1.33402	1.33402	17.60624	0.00852	Significant
Error	5	0.37885	0.07577
Lack of fit	3	0.34945	0.11648	7.92404	0.11412	Not significant
Pure error	2	0.0294	0.0147
Total	14	73.71849
R2 = 0.9948			Adj R2 = 0.9856

. The ANOVA also determined the effects of significant individual terms and their interaction with the selected response [72]. The determination coefficient (R²) and correlation coefficient values reflect the goodness or lack of fit. The statistical significance of the proposed model was analyzed using the F-value, supported by the p-value, within a 95% confidence interval. RSM was used to tune the respective p-values and F-values, which improved the goodness of fit of the model. High p-values (>0.05) of interaction terms were considered insignificant and were therefore excluded from the model. Neglecting these insignificant terms (p > 0.05) enhanced the model’s overall fitness. The significance of each regression coefficient was checked using the p-value, which indicates the probability of error. At the 95% confidence level, the results indicate that the model is highly significant due to the higher Fisher’s test value (F value = 37.07) and lower probability (p-value 0.001), confirming its statistical significance. A p-value of 0.001 indicates that the probability of obtaining a large F-value due to noise is only 0.1% [73]. The model’s fit was also assessed using the coefficient of determination (R²), which indicated how closely the experimental data aligned with the predicted response. Additionally, a higher R² value (0.99) for ester yield suggested that the model was more significant and had better predictive accuracy. According to the ANOVA fit statistics, for a model to fit well, the difference between the predicted R² and adjusted R² should be less than 0.2. The observed difference for the ester yield was significantly lower (0.0092), indicating a better fit.

2.5. Optimization of Reaction Conditions by RSM Analysis of Ester Yield %

2.5.1. Effect of Reaction Time on Upgrading of Rice Straw Bio-Oil in Butanol

The research examined how varying holding times (from 1 to 3 h) affected the upgrading of rice straw bio-oil in butanol at different temperatures, as shown in Table 2. Experiments (1 and 9) at 250 °C, (3 and 13) at 300 °C, and (2 and 5) at 350 °C were conducted under similar conditions, with the only variations being the holding times set at 1 and 3 h for each experiment. Interestingly, the holding times at both extremes had little impact on the yields of oil, aqueous, and solid products. The only significant change observed was a consistent increase in the yield of gaseous products as the holding time extended from 1 to 3 h. Elemental composition and the HHV were compared in Table 3, revealing no significant changes. Table 4 also presented the physical properties of bio-oil and the upgraded experiments (1 and 9) at 250 °C, (3 and 13) at 300 °C, and (2 and 5) at 350 °C. These data indicated that holding times at the same temperature during upgrading under supercritical butanol conditions had minimal to no impact on the physical properties, including water content, density, and viscosity of the upgraded oil. Table 6 outlines the experimental matrix, highlighting the ester yield values. Esters were identified as the predominant functional group of compounds. Comparing the ester yields from upgraded experiments (1 and 9) at 250 °C, (3 and 13) at 300 °C, and (2 and 5) at 350 °C, it was surprising to find that the highest ester yields (44.83%, 44.59%, and 44.68%) were all observed in experiments (4, 14, and 15) with a reaction holding time of 2 h. This observation is consistent with Figure 7, which depicts 2D graphic surface and contour optimization plots of ester yield, indicating that a reaction time of 2 h is optimal. It can be confidently stated that ester formation peaks after 1 h, reaching its maximum at 2 h. Beyond this point, the ester begins to degrade, resulting in a decreased amount after 3 h of reaction time. Figure S3 further illustrates this in a 2D interaction of time, temperature, and catalyst on ester yield.

2.5.2. Effect of Temperature on Upgrading of Rice Straw Bio-Oil in Butanol

The bio-oil upgrading experiments were conducted at temperatures between 250 and 350 °C, with catalyst weights ranging from 0.60 to 1.80 g and reaction times spanning 1 to 3 h, as shown in Table 2. The product yields from supercritical butanol upgrading of rice straw bio-oil are detailed for experiments labeled 1 to 15. Additionally, a control experiment was conducted at the suspected optimal conditions of 300 °C with 1.80 g of the catalyst and a reaction time of 2 h without the catalyst. The oil product yields consistently decreased as temperatures increased from 250 to 350 °C, while gas and aqueous product yields rapidly increased at higher temperatures within the same range. The ZrO₂-Fe₃O₄/AcB catalyst used in the study may facilitate gasification processes at temperatures above 350 °C. Examining the experiments across different temperatures, catalyst weights, and reaction times indicates that higher temperatures contributed to the gasification of reaction products. This resulted in increased gas yields and reduced oil yields, especially noticeable at 350 °C. In contrast, at 300 °C, which is identified as the optimal temperature in Figure 7, significantly higher yields of oil and ester were observed compared with the conditions at 350 °C. These yields were almost equivalent to those obtained at 250 °C. In summary, improving the efficiency of bio-oil can be achieved by raising the reaction temperature to approximately 300 °C. However, operating at temperatures above 350 °C may encourage coke formation, which can dominate and eventually deactivate the catalyst, thereby reducing the ester yield [74].

2.5.3. Effect of Catalyst on Upgrading of Rice Straw Bio-Oil in Butanol

A clear trend is observed when assessing the catalyst’s effect on gasification. This is demonstrated by comparing Experiments 3, 4, 7, 8, 13, 14, and 15 with the control at a constant reaction temperature of 300 °C, despite varying catalyst weights, as shown in Table 2. These experiments consistently resulted in a gas yield that was double that of the control experiment (9.91%). Experiments 3 and 13 show the lowest gas yield with the least catalyst load (0.6 g). In the control experiment, where no catalyst was used, the yield of oil products was significantly higher at 52.04%. This contrasts with the experiments involving a catalyst (Experiments 1–15), where the oil yield ranged from 23.57% to 47.08%. Additionally, the control experiment without a catalyst produced a lower yield of solids at 0.69%, in contrast to the experiments with a catalyst, which yielded between 1.12% and 3.71%. Table 3 shows a significant increase in carbon content and a decrease in oxygen content at a catalyst concentration of 1.8 g, with carbon rising from 54.1% to 66.6% and oxygen dropping from 36.3% to 21.29%. As a result, the esterified bio-oils in experiments using a catalyst weight of 1.8 g exhibited Higher Heating Values (HHVs) that exceeded those of raw bio-oil and experiments with catalyst weights of 0.6 g and 3.0 g. These observations are consistent with the 2D graph of Figure 7, where the catalyst weight is maintained at 1.8 g, as well as with Table 6, which presents the experimental matrix and predicted ester yield percentages. The optimal condition for the highest ester yield was achieved, with an experimental ester yield of 44.83% and a predicted ester yield of 44.7%. These findings clearly show that the inclusion of a catalyst significantly affected the quantities of liquid, solid, and gas outputs and are consistent with previous studies [75].

2.5.4. Proposed Esterification Reaction Mechanism

Figure 8 illustrates the reaction mechanism of Lewis acid-catalyzed esterification reaction. The esterification mechanism typically involves several steps: (1) Acetic acid, present in the bio-oil, adsorbs onto the catalyst surface through an interaction with its carbonyl group; (2) the catalyst polarizes the carbonyl group, forming a carbocation. This facilitates a nucleophilic attack by the oxygen lone pair of butanol; (3) a substitution reaction then occurs, resulting in the formation of the butyl acetate ester and the release of a water molecule [76]. The hydrogenation mechanism often involves two primary steps: (1) the metal oxide catalyst dissociates hydrogen molecules into adsorbed hydrogen atoms, and (2) the unsaturated compounds in the bio-oil chemisorb onto the catalyst surface, where they are subsequently reduced by the adsorbed hydrogen atoms.

2.6. Catalyst’s Stability

The durability of the ZrO₂-Fe₃O₄/AcB magnetic nanoparticle catalyst was evaluated through a catalyst reusability test involving four recycling tests under optimal conditions (300 °C, 2.0 h, and 1.8 g), replicating the conditions of Experiment 4. Figure 9 illustrates the stability of the catalyst over these four runs, with no apparent loss in catalytic activity. The calculated ester yield reduced from 44.83% to 40.5% over four runs, indicating a 4.0% reduction, while the acids increased by approximately 6.0%. This decline in ester yield could be attributed to two main factors. Firstly, the adsorption and aggregation of oligomeric products on the acid sites within the catalyst’s pores may lead to a substantial decrease in its reactivity. Secondly, it could be related to a small amount of metal oxide leaching during the four runs. This observation is consistent with the existence of some empty regions in the used catalyst (Figure 4i,j,l). Importantly, the spent catalyst was successfully employed for subsequent runs under the same optimal reaction conditions after undergoing washing and drying procedures.

3. Materials and Methods

3.1. Materials

Rice straw was obtained from the Delta Research and Extension Center (DREC) in Stoneville, Mississippi. Zirconium (IV) oxychloride octahydrate ZrOCl₂⋅8H₂O (Acros Organics, each with a purity level of > 98%) and Ferric chloride hexahydrate FeCl₃⋅6H₂O were purchased from Sigma-Aldrich with a purity of ≥99%. Sodium hydroxide (NaOH) and potassium bicarbonate (K₂CO₃) with a purity of ≥99% were also purchased from Sigma-Aldrich. Methanol, butanol, and acetone, all with a purity of ≥99%, were procured from Fisher Scientific and used without additional purification.

3.2. Pyrolysis of Rice Straw and Activation of Rice Straw Biochar

The pyrolysis of rice straw was conducted at a feed rate of approximately 7 kg/h in a stainless-steel auger reactor. The detailed methods of pyrolysis and activation of rice straw biochar were described in the previous publication [24].

3.3. Preparation of ZrO₂-Fe₃O₄/AcB

Biochar-based catalysts of ZrO₂-Fe₃O₄/AcB were prepared using the wet impregnation method. The fabrication of the ZrO₂-Fe₃O₄ bimetallic catalyst supported on rice straw biochar (AcB) was performed based on the previously reported method [77]. Likewise, 1 g of rice straw biochar was dispersed in 30 mL of deionized water then the solution was sonicated for 15 min. The weight ratio (1:1:3) of zirconium, iron, and rice straw biochar, respectively, was used to fabricate the catalyst’s metallic composition. Weights of 1 g of ZrOCl₂⋅8H₂O and 1.5 g FeCl₃⋅6H₂O were dissolved in 50 mL of deionized water and sonicated for 15 min. Then, a zirconium-iron salt mixture was gradually added to the rice straw biochar with constant stirring. The suspension was stirred at room temperature for 1 h, then dried in a vacuum oven at 80 °C overnight. Subsequently, the mixture was inserted in OTF-1200X furnace tube (MTI Corp, Hefei, China) and heated to 750 °C under nitrogen flow (1 L/min) with a heating rate of 10 °C/min and settled at 750 for 5 h. The activated biochar (AcB) support made at 800 °C was employed as a support. The overall weight percentage of the catalysts was 100%, while the total weight percentage of bimetallic (Zr:Fe) was 40%. The bimetallic (Zr:Fe) ratio employed in the preparation was (1:1).

3.4. Catalyst Characterization

Characterization of ZrO₂-Fe₃O₄/AcB catalyst was performed using several analytical techniques such as FTIR, TGA, XRD, SEM-EDX, and BET surface area. The thermogravimetric analysis (TGA) was performed by the SDT Q600 TA instrument (Eden Prairie, MN, USA). The analysis was conducted between 20–800 °C at a heating rate of 10 °C/min under an N₂ atmosphere. X-ray diffraction (XRD) patterns of the prepared catalyst were obtained using a Rigaku Smartlab X-ray diffraction system (Rigaku Corp., Akishima-shi, Japan) operated at 40 kV and 44 mA using Cu-Ka radiation with a wavelength of 1.54 Å, from 3° to 90° at a scan rate of 1°/s. Peak intensities were recorded every 0.03° at a sweep rate of 1.0° (2θ/min). The catalyst’s morphology and elemental composition (Zr, Fe, Si, O, and C) were obtained by using JEOL 6500F Field Emission Scanning Electron Microscope with an energy-dispersive X-ray spectroscopy (JOEL, Peabody, MA, USA). Both samples were coated with a 15 nm layer of platinum and imaged at a 5 keV accelerating voltage. A Fourier Transform Infrared (FTIR) spectroscopic analysis was conducted on the ZrO₂-Fe₃O₄/AcB catalyst by using a Perkin Elmer FTIR spectrometer. The analysis covered a wavelength range from 500 to 4000 cm⁻¹. The Brunauer–Emmett–Teller (BET) surface area, pore volume, and pore size distribution (BJH) of AcB and ZrO₂-Fe₃O₄/AcB were determined by analyzing the adsorption and desorption isotherms of nitrogen at a temperature of −196 °C. This analysis was conducted using a Quantachrome Autosorb iQ gas sorption analyzer (Quantachrome, Boynton Beach, FL, USA).

3.5. Bio-Oil Upgrading Process in Butanol

The bio-oil was initially combined with butanol (3:1 w/w), and the mixture was employed for esterification. A total of 100 g of the bio-oil/butanol mixture was taken for the catalytic esterification process in a 450 mL Parr reactor equipped with a temperature controller, stirrer, and heating mantle. Varying weights of the catalysts were taken for the reaction. The weighted catalyst was placed in the reactor with the butanol/bio-oil combination. Then, the reactor was purged with 100 psi H₂ five times to get rid of air inside the reactor. Subsequently, the reactor was pressurized to 1000 psi with H₂ at room temperature. The reactor was heated to the reaction temperature (250–350) °C with a heating rate of 5.0 °C/min. The reaction temperature was kept for 1 to 3 h at a stirring speed of 800 rpm. The reaction conditions of the bio-oil/butanol ratio (3:1) were maintained consistently across all 15 tests. After the completion of each reaction, the heater was switched off, and the reactor was cooled to room temperature using an ice-water bath. The catalyst particles were isolated from the liquid products using an external magnet. The liquid products were left to stand overnight, allowing the immiscible aqueous fraction (AF) and oil fraction (OF) to separate. The recovered solids were rinsed several times with acetone, dried at 105 °C for 12 h, and weighed. Yields of liquids (biofuel and aqueous fraction), gas, and solid products were determined based on calculation using the following equations:

$$Y \left(biofuel\right)={\displaystyle \frac{Mass \left(biofuel\right)}{Mass \left(bio-oil/butanol\right)}}\times 100\%$$

(1)

$$Y \left(Aqueous\right)={\displaystyle \frac{Mass \left(Aqueous\right)}{Mass \left(bio-oil/butanol\right)}}\times 100\%$$

(2)

$$Y \left(Solid\right) ={\displaystyle \frac{Mass \left(Solid\right)}{Mass \left(bio-oil/butanol\right)}}\times 100\%$$

(3)

$$Y \left(Gas\right) ={\displaystyle \frac{Mass \left(gas\right)}{Mass \left(bio-oil/butanol\right)}}\times 100\%$$

(4)

3.6. Physical and Chemical Characterization of Raw and Esterified Bio-Oil

The physical and chemical properties of rice straw bio-oil, such as pH, water content, viscosity, acid value, Higher Heating Values, FTIR, and GC/MS analysis, were performed according to the methods described in the previous study [24].

4. Conclusions

Carboxylic acids emerge as the most abundant components in bio-oil derived from biomass pyrolysis, showcasing the potential of these renewable sources for energy and biobased chemicals. Esterification, particularly the production of butyl esters, is achievable through the esterification process of rice straw bio-oil in supercritical butanol in the presence of H₂ gas. A magnetic bimetallic oxide catalyst supported on activated rice straw biochar (ZrO₂-Fe₃O₄/AcB) was developed and characterized for the esterification process. The key factors influencing the esterification conversion process, such as temperature, time, and catalyst concentration, were studied and optimized. Using an experimental design approach and optimizing through response surface methodology (RSM), the highest yield achieved was 44.83%. The identified optimal conditions for this result, including 300 °C temperature, 1.80 g catalyst loading, and a reaction time of 2.0 h, can serve as a starting point for further exploration of the bio-oil conversion process. However, temperatures exceeding 350 °C may promote coke formation, which can dominate and eventually deactivate the catalyst, leading to a decrease in ester yield. The HHV experienced a significant increase, rising from 21.34 MJ/kg in the raw bio-oil to 33.1 MJ/kg in the upgraded bio-oil. Additionally, there were notable improvements in physical properties, including a 60% reduction in water content and a 20% decrease in TAN, indicating lower acidity. The RSM suggests that the best conditions for achieving high ester content involve exploring temperatures above 295 °C, a duration of 2.3 h, and a catalyst weight of 1.8 g. Finally, the catalyst can be separated by an external magnet and reused for up to four cycles while maintaining its catalytic activity.