Biodiesel Stability Enhancement Through Catalytic Transfer Hydrogenation Using Glycerol as Hydrogen Donor

Abstract

This research aimed to enhance biodiesel stability through catalytic transfer hydrogenation using a biomimetic bimetallic catalyst and glycerol as a hydrogen donor. The effects of catalyst species, intermediate solvent, glycerol feed, and glycerol form on biodiesel stability were investigated. In this study, the examined bimetallic catalysts were Zn-Cr-bicarbonate, Zn-Cr-formate, Zn-Cr-Ni, and Cu-Ni/SiO₂. Based on the results, the most excellent catalyst was presented by Cu-Ni/SiO₂ catalyst with DMF solvent and 10 wt% glycerol feed. This combination demonstrated a significant reduction in iodine (ΔIV = −4.9 g-I₂/100 g) and peroxide values (ΔPV = −5.2 meq-O₂/kg) accompanied by an elevation of oxidative stability (ΔOS = 4.3 h). Moreover, the reaction of catalytic transfer hydrogenation using these bimetallic catalysts followed the theoretical mechanism of the simultaneous dehydrogenation–hydrogenation process with two different metals. The promotion of bicarbonate and formate ions on the bimetallic catalyst provided hydrogen transfer assistance in the catalyst. Hence, the continuous improvement of biodiesel properties is expected to promote sustainable implementation of cleaner diesel fuel.

1. Introduction

In the last decade, the direction of worldwide policy has been focused on the development of renewable energy for energy fulfillment. This program is intensified to reduce dependence on fossil energy, combat global warming, anticipate negative impacts on the economy, and achieve sustainable utilization [1,2]. There are many choices of renewable energy sources that can be promoted, i.e., solar, wind, hydro, geothermal, and biofuels [3]. Among those sources, biodiesel is one of the biofuels that has been produced and used massively as a substitute for petrol diesel. Based on data released by OECD and FAO [4], world biodiesel production reached 60.6 billion liters in 2024, almost twice as high as production in 2015. The three leaders in biodiesel production were the European Union, the United States, and Indonesia. The accumulation of biodiesel production from these three producers was 69% of total global production. This outlook also forecasts that biodiesel production will continue to grow upwards in line with future elevation in energy demand.

Biodiesel offers several advantages over petrol diesel, including reduced exhaust emissions, superior lubricity due to the ability of long-chain FAMEs to form stable lubricating layers on metal surfaces, a higher flash point, improved biodegradability, and lower toxicity [5,6,7]. Additionally, biodiesel can be blended with fossil diesel or even other biofuels, providing significant greenhouse gas reduction benefits [8]. However, its practical application is hindered by challenges such as poor cold flow properties and low oxidative stability [9]. The poor cold flow properties are attributed to a higher concentration of saturated FAMEs, as these compounds are more likely to form gum, potentially causing blockages in vehicle fuel system pipes and filters during cold weather conditions [10]. In contrast, the low oxidative stability is primarily attributed to the presence of polyunsaturated FAMEs, such as methyl linoleate and methyl linolenate, which can degrade into undesirable products like aldehydes, ketones, peroxides, and acids [11,12]. Furthermore, contaminants such as water, metals, and other impurities can further compromise the oxidative stability of biodiesel. Oxidation of biodiesel has been reported to cause saponification, leading to an increase in the total acid number and kinematic viscosity [13]. Both critical challenges must be addressed to enable higher biodiesel blends in fossil diesel.

Palm biodiesel typically has lower oxidative stability compared to fossil diesel, with induction periods ranging from 3–12.5 h, which falls below the minimum standard of 30 h for Category 1 set by the Worldwide Fuel Charter [14,15]. Poor oxidation stability of biodiesel can result in the formation of oxidation products that cause sedimentation, corrosion, and filter blockages in fuel injection systems, leading to potential engine damage and increasing exhaust emissions [11,16]. Several strategies, including antioxidant insertion, blending, fractionation, geometric isomerization, and partial hydrogenation, have been proposed to address oxidative stability issues caused by double bonds in fatty esters. Antioxidant insertion is known as a conventional treatment for biodiesel oxidation stability enhancement [17]. Antioxidant, whether natural or synthetic, is incorporated into biodiesel to inhibit the oxidation of its fatty esters [18,19]. Typically, this additive provides positive effects on cold flow properties [20], NO_x emission [21], and production cost [6]. It prevents crystal growth in cold conditions by the co-crystallization pathway [22]. Nonetheless, this addition has a negative impact on reducing biodiesel kinematic viscosity [6] and calorific value [23].

Later, biodiesel blending with fossil diesel or better oxidation stability biodiesel is proposed to solve this challenge. The use of fossil diesel as a blending agent is a simple and rapid method to significantly improve biodiesel oxidation stability due to the minimum content of double bonds and oxygenates in the fossil diesel [24,25]. However, this treatment results negatively in environmental aspects, such as biodegradability alleviation and SO_x emission elevation [26]. In the future, it is also potentially constrained in terms of sustainability and economics due to non-renewability and high prices, respectively [27]. To address those issues, mixing with higher oxidation stability biodiesel has emerged as an alternative. Specifically, this method utilizes saturated fatty ester-rich biodiesel to refine oxidation stability [26,28]. This mixing agent also promotes a better cetane number for the mixture [28]. Unfortunately, it has poor cold flow properties [9]. Hence, the application of this process requires optimization to minimize the negative effect.

For the third option, fractionation is conducted to separate biodiesel components physically based on the dissimilarity of freezing or boiling points. Winterization and solvent fractionation are demonstrated to fractionate various fatty esters by cooling them at certain temperatures [17]. The solidified components are segregated as products with better oxidation stability [29]. Technically, both processes can be run safely due to the low operating temperature and atmospheric pressure. Another investigated process is vacuum distillation, which utilizes boiling point differences to split biodiesel components. Vacuum condition is chosen to avoid high-temperature processes that may lead to thermal cracking [30]. Although the operating temperature is lower, this process still consumes a lot of energy to provide a vacuum condition. The desirable product, which is higher oxidation stability and lower boiling point, is obtained in the distillate [31]. The main products of those fractionation treatments are categorized as rich in saturated fatty esters, which are associated with higher cetane numbers [32]. Nevertheless, they exhibit low cold flow properties [31,33]. Additionally, another drawback is the reduction in biodiesel yield because some fractions are removed [17].

Hereafter, geometric isomerization also presents progressive performance in enhancing oxidation stability. This process primarily targets the transformation of “cis-” structure into “trans-” structure in unsaturated fatty esters [34]. For instance, this conversion will improve the oxidation stability of C18:1 fatty ester from 2.5 h (methyl oleate) to 7.7 h (methyl elaidate) [35]. Generally, this reaction is run at a temperature range of 100–300 °C and a pressure range of 1–10 bar [36]. These reaction configurations potentially create hazards in operating conditions, mainly at high temperature and pressure. The combination of high temperature and pressure may also result in an energy-intensive and costly process. Moreover, the isomeric products tend to slightly elevate melting and cloud points [35].

Furthermore, partial hydrogenation is the most prominent method for oxidation stability enhancement [37]. This process partially attaches hydrogen to the double bonds, resulting in some of them becoming saturated [17]. In this topic, the intended components for hydrogenation are polyunsaturated fatty esters (PUFEs), and the expected products are mono-unsaturated fatty esters (MUFEs) [38]. This treatment conclusively solves the oxidation stability challenge at the root of the problem [39]. It is represented by the product quality, which presents magnificent oxidation stability and acceptable cold fluidity characteristics [40]. Despite its high energy consumption, this technology is confirmed to have minimal effect on diesel engine operation [41].

According to these superiorities, partial hydrogenation is considered for further development in improving oxidative stability. Chen et al. [8] conducted partial hydrogenation of palm biodiesel using a fixed-bed reactor at 100 °C and 0.3 MPa with catalysts 0.5 wt%-Pd/SBA-15 and 0.5 wt%-Pd/Zr-SBA-15. The study demonstrated a significant improvement in biodiesel oxidative stability, increasing the induction period from 19.4 to 27.9 h, surpassing the oxidative stability requirements of 20 h set by the European standard for B7 in 2013. Ramayeni et al. [42] performed partial hydrogenation at 120 °C and 6 bar using a 5 wt%-Ni/C catalyst, reducing the iodine value of palm biodiesel from 91.78 to 82.38 g-I₂/(100 g-FAME). Sukjit et al. [43] carried out partial hydrogenation of palm biodiesel at approximately 140 °C and 0.5 MPa hydrogen pressure with a palladium-based catalyst, reducing methyl linoleate content from 8.67 wt% to 0.60 wt%. While effective, partial hydrogenation using hydrogen gas at elevated temperature and pressure and expensive catalysts is less feasible for large-scale applications [44]. Moreover, this method is also more harmful because of the explosion characteristic of hydrogen gas [45]. Therefore, more cost-effective and safer alternatives are needed.

One approach to hydrogenate unsaturated FAMEs without special hydrogen gas handling is through catalytic hydrogen transfer. In this method, hydrogen is transferred by a certain catalyst from hydrogen-donating chemicals to targeted FAMEs under moderate operating conditions [46]. There are many chemical options for hydrogen donors, i.e., formic acid, sodium formate, ammonium formate, sodium borohydride, hydrocarbon solvents, and short-chain alcohols [47,48]. Nevertheless, most of them are poisonous, corrosive, and/or evaporative [44]. Then, glycerol is emerging as a potential hydrogen donor due to its high hydrogen content, renewability, biodegradability, non-poisonousness, non-explosiveness, and non-volatility [49,50,51]. As a by-product of biodiesel production, its availability is also abundant [16,52]. Costa et al. [53] reported that 1 kg of crude glycerol was approximately generated in every 10 kg of synthesized biodiesel. However, the overabundance of this resource poses significant challenges for the biodiesel industry, as its utilization remains limited and its purity is still relatively low (≤50%) [54]. Meanwhile, the most glycerol-using processes, such as CTH, require high-purity glycerol (≥95%) to minimize reaction interferences caused by water [55,56]. Thus, the crude glycerol generated from biodiesel production needs to be purified before being used in the CTH process. Overall, leveraging glycerol as a hydrogen source in the biodiesel hydrogenation provides an impressive solution to enhance biodiesel oxidative stability.

Hypothetically, the transfer hydrogenation mechanism of less saturated FAME into more saturated FAME occurred gradually, as illustrated in Figure 1. In the first stage, methyl linolenate feed is hydrogenated by glycerol hydrogen transfer to produce methyl linoleate. Methyl linoleate product and feed (if any) are then saturated into methyl oleate. Lastly, methyl oleate product and feed (if any) are further converted into methyl stearate by hydrogen addition. In the transfer hydrogenation, glycerol tends to transform into DHA as a by-product [57,58]. DHA is well known as a high-value compound for pharmaceutical, cosmeceutical, and food products [59]. Thus, this process is attractive to study further because it offers dual economic and environmental benefits.

In thermodynamic analysis, the hydrogen transfer reaction between unsaturated FAME and glycerol to produce more saturated FAME and DHA is evaluated using Gibbs free energy. Based on the data of standard Gibbs free energy of formation shown in Table 1, the first reaction of methyl linolenate hydrogenation has a Gibbs free energy of −69.88 kJ/mol. The second step for methyl linoleate hydrogenation is calculated to have a Gibbs free energy of −69.88 kJ/mol. Later, the Gibbs free energy of the third stage of the hydrogen transfer reaction is −159.08 kJ/mol. Thus, this three-stage hydrogen transfer reaction can occur under mild conditions.

Generally, this reaction requires metal-based catalysts to take hydrogen out from a donor and direct it towards an acceptor [60]. The noble metal catalysts, such as platinum (Pt), palladium (Pd), gold (Au), ruthenium (Ru), and rhodium (Rh), are most often used for this process because of their high catalytic performance [61,62]. However, the use of these catalysts on a commercial scale is not economical, and their availability is restricted [46,63]. On the other hand, the low-cost metal catalysts, i.e., iron (Fe), copper (Cu), nickel (Ni), chromium (Cr), cobalt (Co), and zinc (Zn), have lower catalytic reactivity [61,64,65,66]. Therefore, it is necessary to develop effective non-precious metal catalysts. One achievement in improving their performance is the application of a dual metal group in a single catalyst [67]. The first metal group is intended for glycerol dehydrogenation. Then, the second group is expected to hydrogenate the targeted compound. The previous studies have explored this catalyst category for transfer hydrogenation, such as Cu-Ni catalyst for methyl levulinate-to-GVL conversion [68], Cu-Ni catalyst for furfural alcohol production [69], Co-Ni catalyst for phenolic synthesis [70], Fe-Ni catalyst for organosolv lignin treatment [71], Zn-Ni/OMC catalyst for levulinic acid-to-GVL reaction [72], Ag-Ni/ZrO₂ catalyst for GVL manufacturing [73], and Co-Cr/CN catalyst for α,β-unsaturated aldehydes transformation [74]. The presence of nickel in the most prepared catalysts indicates the superiority of nickel performance, compared with other non-precious metals, in the hydrogenation process [67]. With these discoveries, the development of dual metal group catalysts is interesting to be seriously advanced.

Table 1. Standard Gibbs free energy (∆Gf°) for the CTH-related compounds (calculated using equation from Dorofeeva et al. [75]).

Compounds	Formula	∆Gf° (kJ/mol)
Methyl linolenate	C17H29COOCH3	115.84
Methyl linoleate	C17H31COOCH3	35.62
Methyl oleate	C17H33COOCH3	−44.60
Methyl stearate	C17H35COOCH3	−124.82
Glycerol	C3H5(OH)3	−438.52
DHA	CH2OHCOCH2OH	−428.18

Table 1. Standard Gibbs free energy (∆Gf°) for the CTH-related compounds (calculated using equation from Dorofeeva et al. [75]).

Compounds	Formula	∆Gf° (kJ/mol)
Methyl linolenate	C17H29COOCH3	115.84
Methyl linoleate	C17H31COOCH3	35.62
Methyl oleate	C17H33COOCH3	−44.60
Methyl stearate	C17H35COOCH3	−124.82
Glycerol	C3H5(OH)3	−438.52
DHA	CH2OHCOCH2OH	−428.18

The proposed advancement is the implementation of a biomimetic mechanism for dual metal group catalysts. This catalyst development is specifically aimed at mimicking the working mechanism of natural hydrogen transfer. In nature, certain microorganisms produce hydrogen with the assistance of the glycerol dehydrogenase enzyme, which anaerobically dehydrogenates glycerol into DHA while generating hydrogen [59]. As explained by Ruzheinikov et al. [76], this enzyme has a zinc ion (Zn²⁺) as a single active metal located in the nuclei, which plays a role in holding the glycerol tightly under alkaline conditions. In this situation, the β-hydrogen and the hydroxyl-contained hydrogen can be released easily for NAD⁺ reduction. For the process without NAD⁺ involvement, the last-mentioned step requires additional active metal(s), other than zinc, to selectively hydrogenate the targeted compounds. It is hypothesized to enhance the catalyst’s performance.

According to the review above, this study principally aimed to enhance the biodiesel stability through catalytic transfer hydrogenation using glycerol as a hydrogen source and biomimetic catalysts as a solid catalyst. The synthesized biomimetic catalysts were Zn-Cr-bicarbonate, Zn-Cr-formate, and Zn-Cr-Ni. The combination of zinc and chromium has not been widely explored to accelerate hydrogen transfer reactions. Zinc (Zn) was designated as a dehydrogenating metal that resembled a dehydrogenation enzyme. Meanwhile, chromium (Cr) was plotted as an active site for the hydrogenation process because of its strong reducing properties. The addition of nickel (Ni) in the third prepared catalyst was considered to improve hydrogenation activity due to its excellence as a reducing agent [77]. Then, the existence of bicarbonate and formate in the catalyst was signified to assist hydrogen transfer internally from the dehydrogenating metal to the hydrogenating ones. This study also examined the conventional catalyst Cu-Ni/SiO₂ as a reference. The catalytic performance of all prepared catalysts was evaluated using oxidative stability, iodine value, and peroxide value. Additionally, the effect of tricalcium octaglyceroxide, as a hydrogen donor replacing glycerol, on biodiesel stability was also assessed for comparison with normal glycerol. Tricalcium octaglyceroxide is a glycerol-based component in the form of an alkaline salt. The employment of this hydrogen donor was intended to establish an alkaline environment while providing glycerol for the glycerol dehydrogenation process. It mimicked the natural process of the glycerol dehydrogenation by glycerol dehydrogenase enzyme.

2. Materials and Methods

2.1. Materials

In this study, the chemicals used for catalyst synthesis were zinc(II) oxide (≥99%), chromium(III) chloride hexahydrate (≥96%), sodium formate (≥99%), sodium bicarbonate (≥99.5%), nickel(II) acetate tetrahydrate (98%), nickel(II) oxide, copper(II) oxide (≥99%), silica gel Davisil Grade 62, formic acid (≥95%), ethanol (≥95%), acetone (≥99.5%), and ammonia solution (32%). The transfer hydrogen reaction utilized n-butanol and DMF as solvents. Calcium hydroxide was also required for tricalcium octaglyceroxide production. Then, the materials used for product analysis were cyclohexane (≥99.5%), acetic acid glacial (100%), citric acid (anhydrous powder), Wijs solution, isooctane (≥99.5%), pure sodium thiosulfate pentahydrate, sodium dodecyl sulfate powder, potassium iodide (≥99.5%), and difenilamine (≥99%). All mentioned chemicals were supplied by Sigma-Aldrich Co., Ltd. (St. Louis, MO, USA). Glycerol (≥95%) was purchased from Ecogreen Oleochemicals Co., Ltd. (Medan, Indonesia). Meanwhile, palm biodiesel were manufactured by Sinarmas Bio Energy Co., Ltd. (Jakarta, Indonesia).

2.2. Methods

2.2.1. Catalyst Preparation

Zn-Cr-bicarbonate and Zn-Cr-formate catalyst preparation were conducted through multiple work stages, involving Zn-Cr-Cl preparation, Zn-Cr-bicarbonate synthesis, and Zn-Cr-formate manufacturing. The first stage, Zn-Cr-Cl preparation, was carried out with procedure published by Boehm et al. [78]. Zinc(II) oxide and 0.5 M chromium(III) chloride solution, with a molar ratio of Zn:Cr = 3:1, were reacted in distilled water at 60 °C. It was then followed by filtration, washing, and overnight drying at 110 °C. For the next stage, Zn-Cr-Cl solid was further treated with sodium carbonate and sodium formate solutions at 60 °C to produce Zn-Cr-bicarbonate and Zn-Cr-formate, respectively.

Zn-Cr-Ni catalyst was synthesized by modifying Zn-Cr-bicarbonate or Zn-Cr-formate materials. The first step was nickel formate dihydrate provision. It was prepared by reacting nickel(II) acetate tetrahydrate, formic acid, and ethanol. The microcrystalline precipitate was washed with ethanol three times and acetone once. The washed solid was then dried overnight in the oven. The second step was Zn-Cr solid synthesis by heating Zn-Cr-bicarbonate or Zn-Cr-formate at 300 °C for 6 h. The last step was nickel impregnation on Zn-Cr solid. This process was conducted by dripping nickel formate solution in ammonia (containing 1.0 g Ni formate) on 3.0 g Zn-Cr solid (incipient wetness impregnation method). The impregnated solid was then dried overnight at 110 °C in the oven.

Afterwards, Cu-Ni/SiO₂ catalyst was prepared by dripping Cu-Ni solution on silica gel Davisil Grade 62 (incipient wetness impregnation method) [79]. Cu-Ni solution was made by mixing copper(II) oxide and nickel(II) oxide, with a molar ratio of Cu:Ni = 3:5, in a formic acid/ammonia solution. Finally, the impregnated solid was oven-dried at 110 °C to constant weight.

2.2.2. Catalytic Transfer Hydrogenation

The experimental apparatus setup, as depicted in Figure 2, consisted of a 250 mL three-neck reactor equipped with a reflux condenser, thermometer, glass washing bottle, and under nitrogen flow. Nitrogen flow, during the experiment, was necessary to prevent oxygen exposure in the reactor. The experiment was conducted by mixing 85 g of palm biodiesel, certain glycerol (4 and 10 wt% of biodiesel feed), 4.5 mL of solvent, and 1 g of catalyst in the reactor at an agitation speed of 250 rpm for 6 h. For the reaction with n-butanol and DMF solvent, the temperatures were set at 115 °C and 150 °C, respectively, to avoid solvent evaporation. Later, the temperature was adjusted to 200 °C for a solvent-free reaction. After the reaction, the biodiesel was filtered from the catalyst, washed with distilled water, and then dried in a desiccator for subsequent analysis.

In addition, the effect of glycerol substitution with tricalcium octaglyceroxide as a hydrogen source was also investigated. Tricalcium octaglyceroxide was glycerol in an alkaline solution of calcium hydroxide. It was prepared by homogenizing glycerol and calcium hydroxide with a mass ratio of 1:100. The mixture was then heated at 140 °C for 3 h. Moreover, the demetalling treatment was also assessed to prove the presence of metal catalyst residue in the biodiesel product and its effect on oxidative stability. According to Thiagarajan and Tang [80], this treatment was carried out by washing the biodiesel product with 1 wt% citric acid at a volumetric ratio of 1:1 and a temperature of 60 °C. Later, the treated sample was washed with distilled water (volumetric ratio of 1:1) to remove residual citric acid and metal sites.

2.2.3. Product Analysis

Several analytical methods were assessed for biodiesel feeds and products: oxidative stability, iodine value, and peroxide value. Oxidative stability was measured using the modified Rancimat method (EN 15751). This method determined the induction period—defined as the time elapsed from initial air exposure until the onset of significant oxidative degradation—during accelerated oxidation at 110 °C. Next, for iodine value measurement, the mixture of biodiesel sample, cyclohexane, acetic acid, potassium iodide, and Wijs indicator was titrated with standardized sodium thiosulfate solution. Hereafter, peroxide value analysis was carried out by titrating the biodiesel mixture with standardized sodium thiosulfate solution. In this method, the biodiesel mixture was made by blending biodiesel sample, isooctane, acetic acid, potassium iodide, sodium lauryl sulfate, and starch solution (indicator). These three parameters were presented in the form of their changes (Δ), where the calculation generally followed Equation (1).

Δ = Product value − Feed value

(1)

3. Results and Discussion

3.1. Catalyst Performance Results

3.1.1. Effect of Catalyst Species

In this section, all four synthesized catalysts were evaluated on catalyzing the transfer hydrogenation of biodiesel with a glycerol feed of 10 wt% and DMF as a solvent. The results of catalyst performance on biodiesel stability are summarized in

Table 2. Effect of catalyst species on hydrogenated biodiesel stability (reaction temp. = 150 °C; glycerol feed = 10 wt%; DMF as solvent).

Catalysts	ΔIV (g-I2/100 g)	ΔPV (meq-O2/kg)	ΔOS (h)
Zn-Cr-bicarbonate	−3.8 ± 0.15	−4.4 ± 0.20	−9.9
Zn-Cr-formate	−2.1 ± 0.05	−9.9 ± 0.05	−4.9
Zn-Cr-Ni	0.2 ± 0.05	2.3 ± 0.21	−3.7
Cu-Ni/SiO2	−4.9 ± 0.40	−5.2 ± 0.03	4.3

. Based on Table 2, Zn-Cr-bicarbonate and Zn-Cr-formate catalysts behaved positively in reducing iodine and peroxide values. The combination of iodine and peroxide values reduction indicated the alleviation of double bond in biodiesel due to hydrogenation. However, the sample of biodiesel product exhibited lower oxidative stability than the feed. This instability was potentially attributed to the left-over catalyst in the product. Even though its presence in biodiesel products was in small quantities, the metal components mainly promoted oxidation in fatty esters. In addition, the presence of bicarbonate and formate components in the catalyst residue may slightly encourage the oxidation process earlier [81].

Not only due to bicarbonate and formate, the alleviation of oxidative stability was also due to the remaining zinc and chromium in the biodiesel product. It was proven by the increase in oxidative stability after demetalling treatment on biodiesel products, as shown in

Table 3. Effect of demetalling treatment on biodiesel oxidative stability.

Catalyst	Sample	ΔOS (h)
Zn-Cr-bicarbonate	1	0.6
	2	3.2

. By washing with 1 wt% citric acid, the oxidative stability of treated biodiesel was improved by 0.6–3.2 h higher than the unwashed sample. These results confirmed that the lower oxidative stability of the biodiesel product was partly due to residual catalyst sites.

These residual metals could act as an oxidizing agent to oxidize fatty esters, mainly for the unsaturated components. Referring to the previous study conducted by Zuleta et al. [82], the three-stage process of lipid oxidation illustrated the abstraction of hydrogen atoms forming free radical substances and their replacement with oxygen atom(s) in the presence of metal catalysts. Finally, the result was the formation of various hydroperoxide substances (ROOH or ROOR’). The catalytic mechanism of this process is visualized in detail in Figure 3.

Furthermore, the Zn-Cr-Ni catalyst had an impact on increasing iodine and peroxide values in the biodiesel product. Although not significant, the elevation of both values represented the formation of additional double bonds and peroxide substances, respectively. The promotion of nickel on Zn-Cr solid did not perform in improving the catalyst performance for this reaction. It weakened the catalyst’s ability in hydrogenating biodiesel and tended to direct the oxidation process, producing peroxides(Figure 3) such as lipid oxidation introduced by Zuleta et al. [82]. Consequently, the product demonstrated poorer oxidative stability than the feed. Therefore, Zn-Cr-Ni was concluded to be an ineffective catalyst in facilitating the hydrogen transfer reaction.

And then, the Cu-Ni/SiO₂ catalyst demonstrated the most promising performance, showing significant alleviation in iodine (ΔIV = −4.9 ± 0.40 g-I₂/100 g) and peroxide values (ΔPV = −5.2 ± 0.03 meq-O₂/kg). These results denoted an increase in saturation and a decrease in peroxide content, respectively. In line with this observation, the oxidative stability of the biodiesel product was also improved. The combination of copper and nickel provided a mechanism for hydrogen transfer involving a dehydrogenation–hydrogenation process. It also implied that the negative effect of the left-over metals on oxidation stability was reduced due to the improvement in fatty ester saturation.

The subtraction in iodine values for all observed catalysts in this work tends to be lower than prior studies. Zhang et al. [46] reported the performance of the Ni-La-B catalyst for hydrogenating linoleic-rich fatty acid methyl ester with a reduction in iodine value of 65.3 g-I₂/100 g or ~43% of the feed value. This reaction was executed using sodium borohydride as a hydrogen source at a temperature of 35 °C, atmospheric pressure, a catalyst dosage of 10 wt%, and under ultrasonic assistance. In this condition, the elevation in biodiesel oxidation stability was recorded at around 4.36 h. This alteration in oxidation stability was comparable to the result of the Cu-Ni/SiO₂ catalyst in this research (ΔOS = 4.3 h). Later, Xin et al. [48] successfully hydrogenated highly unsaturated fatty esters using a lanthanum-doped Ni-B catalyst with an alleviation in iodine value of 75.8 g-I₂/100 g (~50% of the initial value). Consequently, it impressively enhanced oxidation stability from 4.1 to 35.3 h. This performance was conventionally observed at a temperature of 85 °C, a catalyst loading of 10 wt%, and a reaction time of 2.5 h. Another study conducted by Gao et al. [83] investigated the activity of Ni-Cr-B-CTAB catalyst for fatty ester hydrogenation. This reaction was performed using iso-propanol as a hydrogen source at a temperature of 85 °C, a catalyst loading of 10 wt%, and under microwave assistance for 1.5 h. This reaction setup led to a decline in iodine value of 60 g-I₂/100 g, along with an improvement in oxidation stability by 5 h. The observed change in oxidation stability was similar to this study with the Cu-Ni/SiO₂ catalyst.

3.1.2. Effect of Intermediate Solvent

The effect of intermediate solvent on the catalytic transfer hydrogenation of biodiesel was evaluated using three solvent variations: DMF, n-butanol, and solvent-free. The comparison of these solvent variations is tabulated in

Table 4. Effect of intermediate solvent on hydrogenated biodiesel stability (glycerol feed = 10 wt%).

Catalysts	Solvent	ΔIV (g-I2/100 g)	ΔPV (meq-O2/kg)
Zn-Cr-bicarbonate	DMF	−3.8 ± 0.15	−3.5 ± 0.20
	n-butanol	−4.1 ± 0.20	−1.7 ± 0.25
	No solvent	−1.5 ± 0.20	14.0 ± 0.30
Zn-Cr-formate	DMF	−2.1 ± 0.05	−9.9 ± 0.05
	n-butanol	0.1 ± 0.10	−8.8 ± 0.03
	No solvent	−5.4 ± 0.95	−1.0 ± 0.18
Zn-Cr-Ni	DMF	0.2 ± 0.05	0.0 ± 0.20
	n-butanol	−0.4 ± 0.01	2.3 ± 0.20
	No solvent	−0.3 ± 0.05	0.3 ± 0.06

. For Zn-Cr-bicarbonate and Zn-Cr-formate catalysts, the hydrogen transfer worked more effectively with DMF solvent. The reaction with the Zn-Cr-bicarbonate catalyst successfully reduced iodine and peroxide values by 3.8 ± 0.15 g-I₂/100 g and 3.5 ± 0.20 meq-O₂/kg, respectively. The reduction in iodine and peroxide values for the Zn-Cr-formate catalyst was recorded at 2.1 ± 0.05 g-I₂/100 g and 9.9 ± 0.05 meq-O₂/kg. The implementation of n-butanol solvent and solvent-free for the Zn-Cr-bicarbonate catalyst specifically degraded iodine value, indicating hydrogenation has occurred. However, the opposite changes of peroxide values happened for the solvent-free reaction. This trend led to the formation of additional peroxide substances during the reaction. For the Zn-Cr-formate catalyst, n-butanol solvent only performed in reducing peroxide value. Then, solvent-free transfer hydrogenation was recorded to significantly alleviate iodine value. Meanwhile, the Zn-Cr-Ni catalyst exhibited minimal catalytic activities in the hydrogen transfer reaction across all solvent variations. The reduction in iodine values was insignificant. Moreover, the increase in peroxide values signified that this catalyst tended to encourage the formation of undesired peroxide compounds.

Compared to previous studies, the lowering in iodine values observed in the present investigation was not particularly remarkable. Zhang et al. [84] examined catalytic transfer hydrogenation for Jatropha biodiesel using Raney-Ni catalyst and iso-propanol as a hydrogen donor. The optimum decline in iodine value was 26.9 g-I₂/100 g. In this observation, the hydrogenation was carried out with water as a solvent to facilitate the reaction between the involved components. However, the quantity of water was restricted to prevent its adverse effects in the form of over-dilution and insufficient reaction. Xin and coworkers [48] declared their findings on the effect of water as a solvent on the transfer hydrogenation of double bond-rich fatty esters. The best result was demonstrated by a maximum reduction in iodine value of 81.5 g-I₂/100 g. As a solvent, water was required in the reaction to hydrolyze sodium borohydride and produce hydrogen. Later, this hydrogen was transferred to fatty esters with the support of a metal catalyst.

3.1.3. Effect of Glycerol Quantity

In this segment, the effect of glycerol quantity on biodiesel stability was assessed by applying two variations: 4 and 10 wt% of biodiesel feed. Stoichiometrically, the required glycerol for palm-biodiesel hydrogenation was estimated to be approximately 3.5 wt%. Hence, all variations were set in the condition of excess glycerol. The experimental results of excess glycerol variations are recapitulated in

Table 5. Effect of glycerol quantity on hydrogenated biodiesel stability (reaction temp. = 200 °C; solvent-free reaction).

Catalysts	Glycerol Feed (wt%)	ΔIV (g-I2/100 g)	ΔPV (meq-O2/kg)	Phase (1 atm, 25 °C)
Zn-Cr-bicarbonate	4	0.5 ± 0.15	1.7 ± 0.28	Liquid
	10	−1.5 ± 0.20	11.9 ± 0.30	Partly solid
Zn-Cr-formate	4	−0.3 ± 0.0	1.2 ± 0.51	Liquid
	10	−6.7 ± 0.95	−1.0 ± 0.18	Partly solid

. The higher quantity of glycerol led to the greater reduction in biodiesel iodine value. Then, the best result was presented by reaction with a glycerol feed of 10 wt% and a Zn-Cr-formate catalyst. The decrease in iodine and peroxide values was obtained at 6.7 ± 0.95 g-I₂/100 g and 1.0 ± 0.18 meq-O₂/kg, respectively. In addition, this biodiesel product also demonstrated a partially solid state at atmospheric room temperature. The reduction in iodine value indicated the saturation of fatty esters, including the potential formation of saturated fatty esters. At room temperature, the saturated fatty esters with more than 17 carbon atoms will be solidified due to the higher melting points than room temperature [9]. Therefore, the more alleviation in iodine value potentially generates the more visible solidified components at room temperature. The probability of solid particle formation will also correspondingly elevate as the composition of fatty esters with 18 or more carbon atoms rises.

For comparison, Lu et al. [61] reported their discovery regarding the influence of hydrogen donor amount on the subtraction in iodine value and linoleic conversion ratio. The reaction was conducted by hydrogenating high-linoleic fatty ester with hydrogen donor ammonium formate in the presence of Pd/OB catalyst. The optimum IV reduction was 55 g-I₂/100 g, obtained at a hydrogen donor amount of 40 g (fatty ester feed = 3 g). The excessive amount of hydrogen donor led to over-adsorption of hydrogen donor molecules at the catalyst active sites, hindering the binding of unsaturated fatty esters. Subsequently, Xin et al. [48] investigated several quantities of hydrogen donor to determine its effect on fatty ester saturation. The highest IV reduction was recorded at 76.8 g-I₂/100 g, observed at a sodium borohydride amount of 1.14 g (fatty ester feed = 7 g). In this setup, the hydrogen donor was required in excess amounts to promote higher conversion.

3.1.4. Effect of Different Glycerol Form as Hydrogen Source

This section studies two different forms of glycerol, normal glycerol and tricalcium octaglyceroxide, on transfer hydrogenation. The results of this comparison are shown in

Table 6. Effect of different glycerol forms on hydrogenated biodiesel stability (reaction temp. = 115 °C; hydrogen source feed = 10 wt%; n-butanol as solvent).

Catalysts	Glycerol Form	ΔIV (g-I2/100 g)	ΔPV (meq-O2/kg)
Zn-Cr-bicarbonate	glycerol	−4.1 ± 0.20	−1.7 ± 0.25
	tricalcium octaglyceroxide	−1.7 ± 0.40	−1.7 ± 0.50
Zn-Cr-formate	glycerol	0.1 ± 0.10	−8.8 ± 0.03
	tricalcium octaglyceroxide	−6.2 ± 0.20	−10.5 ± 0.05

. For the Zn-Cr-bicarbonate catalyst, the employment of normal glycerol as a hydrogen donor performed slightly better reduction in iodine values than tricalcium octaglyceroxide. In this setup, the iodine values for glycerol and tricalcium octaglyceroxide investigations experienced a subtraction of up to 4.1 ± 0.20 and 1.7 ± 0.25 g-I₂/100 g, respectively. Hereinafter, the combination of Zn-Cr-formate catalyst and tricalcium octaglyceroxide presented the best performance in reducing iodine and peroxide values. In this reaction, the iodine and peroxide values were alleviated by 6.2 ± 0.20 g-I₂/100 g and 10.5 ± 0.05 meq-O₂/kg, respectively. The utilization of tricalcium octaglyceroxide was intended to boost the alkalinity of the reaction environment. Therefore, the hydrogen transfer reaction was obviously accelerated in an alkaline environment and more potential hydrogen donor molecules.

For comparison, Sancheti and Gogate [85] performed Pd/C catalyst for hydrogenating soybean oil using various formate-based hydrogen donors. The best enhancement (ΔIV = −40.0 g-I₂/100 g) was demonstrated in reaction with hydrogen donor ammonium formate. It was due to the promotion of a stronger alkaline environment and higher hydrogen transfer. Meanwhile, Lu et al. [86] reported on the transfer hydrogenation of corn and soybean oils using glycerol as a hydrogen donor. This observation also utilized iridium (Ir), iron (Fe), and their combinations as catalysts. The highest hydrogenation percentage (95%) was obtained from the reaction with a full iridium catalyst. Then, the lowest performance (7%) was presented by the full iron catalyst. The best metal combination was observed at a molar ratio of Ir:Fe = 1:10, resulting hydrogenation percentage of 90%.

3.1.5. Comparative Study with Previous Findings

Several researchers reported their work on the catalytic transfer hydrogenation of fatty ester and fatty acid feeds, as tabulated in

Table 7. Comparison of hydrogenated fatty esters and fatty acids under different process configurations.

Feed	Catalyst	Hydrogen Donor	Solvent	Temp. (°C)	ΔIV (g-I2/100 g)	ΔOS (h)	Ref
Cottonseed biodiesel	Reney-Ni	Isopropyl alcohol	H2O	80	−31.1	n.a.	[47]
Jatropha biodiesel	Raney-Ni	Isopropyl alcohol	H2O	70–86	−26.9	39.2	[84]
Jatropha biodiesel	Raney-Ni	Isopropyl alcohol	H2O	85	−24.4	n.a.	[87]
Highly unsaturated FAMEs	Ni-La-B	NaBH4	H2O	35	−65.3	4.36	[46]
Highly unsaturated FAMEs	Ni-La-B	NaBH4	H2O	85	−75.8	31.2	[48]
Highly unsaturated FAMEs	Ni-La-B/OB	NaBH4	H2O	75	−62.7	n.a.	[88]
Highly unsaturated FAMEs	Ni-Cr-B-CTAB	Isopropyl alcohol	H2O	85	−60.0	5.0	[83]
Highly unsaturated FAMEs	Pd/OB	Ammonium formate	H2O	80	−57.9	n.a.	[61]
Soybean oil	Pd/C	Ammonium formate	H2O	30	−40.0	n.a.	[85]
	Pd/C	Sodium formate	H2O	30	−33.0	n.a.	[85]
	Pd/C	Potassium formate	H2O	30	−26.0	n.a.	[85]
	Pd/C	Formic acid	H2O	30	−25.0	n.a.	[85]
Corn oil	Ir	Glycerol	No solvent	120	n.a. (%-H = 95%)	n.a.	[86]
	Ir-Fe	Glycerol	No solvent	120	n.a. (%-H = 90%)	n.a.	[86]
Palm biodiesel	Cu-Ni/SiO2	Glycerol	DMF	150	−4.9 ± 0.40	4.3	This study
	Zn-Cr-formate	Glycerol	n-butanol	115	−6.2 ± 0.20	n.a.	This study

Note: %-H = hydrogenation percentage.

. Generally, catalytic transfer hydrogenation is evaluated according to two primary points: catalyst and hydrogen donor. For the catalyst aspect, metal-based catalysts are still the mainstay for accelerating the transfer hydrogenation process. Based on Table 7, precious metals, such as palladium and nickel, are the main possibilities for catalyst active sites due to their excellent reaction performance. More specifically, the development of nickel-based catalysts has recently begun to be intensified for the CTH process because of their affordable price for commercial utilization [46]. Almost all recent investigations, including this study, conducted transfer hydrogenation with this catalyst type. The most superior results of nickel-based catalyst were reported by Xin et al. [48], reducing iodine number by 75.8 g-I₂/100 g and elevating oxidation stability by 31.2 h. These results were better than this study (Cu-Ni/SiO₂ catalyst), which only alleviated iodine value by 4.9 g-I₂/100 g and improved oxidation stability by 4.3 h. Nevertheless, the best oxidation stability enhancement in this study is still comparable to other studies conducted by Zhang et al. [46] and Gao et al. [83].

Furthermore, the most hydrogen donors used in catalytic transfer hydrogenation are formate-based compounds, sodium borohydride, and β-hydrogen-containing alcohols. Ammonium formate is used more frequently than other formate-based compounds because it provides a stronger alkaline environment and promotes more hydrogen transfer [85]. The highest iodine value reduction from ammonium formate utilization was reported by Lu et al. [61] at 57.9 g-I₂/100 g. Sodium borohydride also presented significant results in lowering iodine value [46,48,88]. Then, glycerol and isopropyl alcohol serve as viable alternatives for hydrogen donors. Currently, these β-hydrogen-containing alcohols, mainly glycerol, have great potential to be massive hydrogen donors due to their abundance. According to Table 7, the subtraction in iodine value with isopropyl alcohol is still higher than with glycerol, including this study. It requires further development to improve catalyst activity and selectivity in transferring hydrogen from glycerol to the target components, although there is a Gibbs free energy barrier for hydrogen transfer [89].

3.2. Theoretical Reaction Mechanism of Catalytic Transfer Hydrogenation

According to the results for all catalysts, the hydrogen transfer reaction has a theoretical mechanism that generally consists of a simultaneous dehydrogenation–hydrogenation process. However, the catalytic activity of a specific catalyst can be approached by a specific reaction mechanism as well. For Zn-Cr-bicarbonate and Zn-Cr-formate, the proposed reaction mechanism of transfer hydrogenation is representatively illustrated in Figure 4. The first stage is glycerol binding to zinc metal and unsaturated fatty ester binding to the chromium site. Later, the zinc metal attracts hydrogen atoms from glycerol to react with hydroxides, forming water on the metal surface. In this part, glycerol is converted into DHA. In the other site, the chromium metal opens the double bond contained in unsaturated fatty esters. Furthermore, the hydrogen atom is sequentially transferred to the formate/bicarbonate and then to the fatty ester, producing more saturated fatty esters. In addition to normal glycerol, this mechanism also applies to glycerol-based alkaline salts, such as tricalcium octaglyceroxide. In the reactor, glycerol-based alkaline salts will react with water to form standard glycerol, which serves as a hydrogen donor, along with an alkaline solution for reaction conditioning.

Afterwards, the theoretical reaction mechanism using the Cu-Ni/SiO₂ catalyst, as shown in Figure 5, is simpler than the two previous catalysts. Copper is responsible for binding glycerol, and nickel is tasked with binding unsaturated fatty esters. Subsequently, copper metal carries hydrogen atoms out of glycerol and transfers them to the nickel site [69]. Before receiving a hydrogen atom, nickel plays a role in opening the double bond of fatty esters [69]. After receiving hydrogen, nickel metal reduces fatty ester by hydrogen addition. Finally, the more saturated fatty ester is released.

According to Figure 4 and Figure 5, the differences between both theoretical reaction mechanisms are visualized more clearly via flowcharts shown in Figure 6. The first difference is pointed out in the step of hydrogen atom attraction. For the conventional CTH mechanism (Figure 6A), hydrogen is attracted to the metal surface in the form of metal–hydrogen bonding (M–H bonding) [83]. In the biomimetic mechanism (Figure 6B), hydrogen is bound and reacted with the hydroxyl group on the metal surface to form water. Theoretically, the reaction-based bonding may attract more hydrogen and faster than the M–H bonding. The water formation reaction occurs spontaneously, so the hydrogen concentration on the metal surface will always approach zero. This situation will encourage faster hydrogen mass transfer from the liquid bulk to the metal surface. This is advantageous in terms of providing hydrogen to hydrogenate unsaturated hydrocarbons. Unfortunately, the generated water may be released at high reaction temperature, before the hydrogen is transferred to the other metal.

The second difference is denoted in the step of hydrogen atom transfer. Conventionally, hydrogen atoms are transferred from metal 1 to metal 2, driven by the distinction in hydrogen concentration. Metal 1, which attracts hydrogen atoms from hydrogen donors (H-donor), has a higher hydrogen concentration than metal 2. Thus, the hydrogen atom is moved from higher to lower hydrogen concentration. For a biomimetic mechanism, bicarbonate or formate is incorporated into the metal catalyst to direct the movement of hydrogen. Bicarbonate or format attracts hydrogen atoms from water and returns hydroxyl to metal 1. Later, the hydrogen atom is transferred to metal 2 for double bond hydrogenation. Hypothetically, a hydrogen atom will move more quickly with the help of a hydrogen-transfer compound. Still, this process has a limitation if the generated water is released faster than the hydrogen transfer rate directed by the hydrogen-transfer compound. Therefore, the hydrogenation process with a biomimetic mechanism is recommended to be carried out at a lower temperature, accompanied by yield optimization.

3.3. Future Research Directions

For the continued advancement of this technology, it is advisable to pursue further steps in the development of biomimetic metal catalysts. The key objective is the enhancement of catalyst activity through biomimetic mechanisms at low or mild temperatures for biodiesel oxidation stability improvement. To achieve this goal, several work actions are required, including: (i) exploration of more effective metal combinations; (ii) investigation of more powerful hydrogen-transfer promoters; (iii) advancement in reaction mechanisms; and (iv) evaluation of long-term oxidation stability. The first expected action is recommended to formulate combinations of non-noble metals, preferably in dual or multiple forms, for more reactive catalysts. This suggestion additionally encompasses the characterization of catalysts, both before and after the reaction. In addition to partial hydrogenation, the metal catalysts are then suggested to simultaneously accelerate cis-trans isomerization, further boosting oxidation stability. Beyond the metal combination, it is imperative to investigate hydrogen-transfer promoters for catalyst intensification in internal hydrogen movement. Furthermore, it is recommended to examine a comprehensive reaction mechanism involving reaction variable optimization, in situ product characterization, reaction kinetics-thermodynamics study, and potential autoxidation evaluation. Finally, the long-term oxidation stability of the biodiesel product resulting from this treatment is also advisable to be evaluated systematically.

4. Conclusions

This study confirmed that catalytic transfer hydrogenation occurred in palm-biodiesel using glycerol as a hydrogen donor. It was indicated by the reduction in iodine and peroxide values for Zn-Cr-bicarbonate, Zn-Cr-formate, and Cu-Ni/SiO₂ catalysts. The most effective catalyst was denoted by the Cu-Ni/SiO₂ catalyst with DMF insertion as solvent and 10 wt% glycerol feed. This combination demonstrated a significant reduction in iodine and peroxide values accompanied by an elevation of oxidative stability. For Zn-Cr-bicarbonate and Zn-Cr-formate catalysts, the reduction in iodine and peroxide values was generally not attended by an increase in oxidative stability. It was due to the residual metals, bicarbonate, and formate in the biodiesel product, triggering oxidation earlier. Hereinafter, the substitution of normal glycerol with alkaline glycerol (tricalcium octaglyceroxide) for Zn-Cr-formate had a significant effect on reducing iodine and peroxide values. Additionally, the reaction of transfer hydrogenation using these bimetallic catalysts followed the theoretical mechanism of simultaneous dehydrogenation–hydrogenation with two different metals. Specific to Zn-Cr-bicarbonate and Zn-Cr-formate catalysts, the presence of bicarbonate and formate served to assist hydrogen transfer internally in the catalyst.