Catalytic Applications in the Production of Hydrotreated Vegetable Oil (HVO) as a Renewable Fuel: A Review

Abstract

The significance and challenges of hydrotreatment processes for vegetable oils have recently become apparent, encompassing various reactions like decarbonylation, decarboxylation, and hydrogenation. Heterogeneous noble or transition metal catalysts play a crucial role in these reactions, offering high selectivity in removing oxygen and yielding desired hydrocarbons. Notably, both sulphided and non-sulphided catalysts exhibit effectiveness, with the latter gaining attention due to health and toxicity concerns associated with sulphiding agents. Nickel-based catalysts, such as NiP and NiC, demonstrate specific properties and tendencies in deoxygenation reactions, while palladium supported on activated carbon catalysts shows superior activity in hydrodeoxygenation. Comparisons between the performances of different catalysts in various hydrotreatment processes underscore the need for tailored approaches. Transition metal phosphides (TMP) emerge as promising catalysts due to their cost-effectiveness and environmental friendliness. Ultimately, there is an ongoing pursuit of efficient catalysts and the importance of further advancements in catalysis for the future of vegetable oil hydrotreatment.

1. Introduction

In today’s technology-driven world, diesel fuels have become crucial for powering various human activities such as manufacturing and transportation. They are considered indispensable for maintaining a high standard of living and are essential for creating business opportunities. However, according to the United States Environmental Protection Agency [1], diesel emissions have significant adverse effects on human health, the environment, climate, and environmental justice. Additionally, concerns about the depletion of non-renewable fossil fuels have prompted global efforts to develop renewable diesel fuel to achieve sustainability goals. Sonnichsen’s data indicate a substantial increase in worldwide biodiesel production, reaching 1677 thousand barrels per day in 2020, which is a 90% increase from 2008 [2]. Both biodiesel and hydrotreated vegetable oils (HVO) are considered biobased fuels [3]. However, HVO stands out due to its more efficient burning, lower emissions, and superior cold and storage properties compared to conventional fatty acid methyl ester (FAME) biodiesel and other biofuels [4].

Hydrotreated vegetable oils, also known as “renewable diesel”, are derived from biomass-based triglycerides, such as vegetable oils, through a process called catalytic hydrodeoxygenation, which primarily produces non-oxygenated aliphatic compounds. This terminology and production process have been extensively reviewed by Knothe [5].

Overall, traditional biodiesel from vegetable oils, which consists of n-alkyl esters, typically methyl esters, of fatty acids, has been used to a lesser extent as fuel in their pure form [6,7], but whether used neat or blended with diesel fuel, significantly reduces the emissions of particulate matter (PM), although its impact on nitrogen oxides (NOx) emissions varies, typically resulting in a minor increase [8]. Biodiesel also tends to decrease the emissions of polycyclic aromatic hydrocarbons (PAHs) [9,10], although the effects on PAH emissions are not always consistent [10,11]. However, this conventional biodiesel may present challenges related to cold flow properties and fuel oxidation stability [4,12].

In this manner, HVO offers several advantages over biodiesel, including broader feedstock options and the ability to be fully substituted for diesel or blended in any ratio without requiring modifications to vehicle fuel systems [13]. In contrast, FAME biodiesel is limited to a maximum blending amount of 5% and may require modifications to fuel systems. Additionally, biodiesel is more prone to issues such as fuel filter blocking and lower stability during long-term storage compared to HVO.

HVO also exhibits lower moisture absorption and higher lubricity when mixed with diesel fuel, meeting the EN standard for preventing fuel system seizure [12,14]. Investments in research and development for HVO are deemed beneficial, as they can lead to reductions in exhaust emissions and fuel consumption [15]. According to Volvo Penta and Boonrod et al., using HVO may cut fossil fuel CO₂ emissions by 90% [16,17]. Despite these benefits, there is still little study being conducted on HVO production; much of it focuses on the fuel’s remarkable qualities and performance, which make it a viable alternative to regular diesel. In order to offset the low density and high cetane number of HVO, Aatola also recommends tuning engines for high-concentration blends [12]. He does, however, affirm that the benefits of reduced NOx and PM are still realised while using the original fuel injection timing. Knothe claims that adding branched chain alkanes can improve cold flow characteristics and that HVO can provide cloud points that are on par with diesel fuel [5]. Nylund states that tidy HVO may be used successfully in a fleet of buses without requiring any engine alterations [18]. The fact that HVO uses a similar feedstock to biodiesel and that its cost is now far greater than that of both diesel fuel and biodiesel are two major drawbacks, though. Taking into account the aforementioned information, researchers have faced difficulty due to the consistent rise in demand for biodiesel caused by legal mandates requiring the inclusion of biodiesel blends in transportation fuels in some nations. Their job is to create more sophisticated and efficient processes for turning vegetable oils into biodiesel.

In this work, an unambiguous overview of different catalytic systems is illustrated to determine the optimal reaction conditions for the hydrotreatment process. The various heterogeneous catalysts as transition metals and their oxides on numerous supports are demonstrated to compare each of their performance in the actual process. In addition to that, noble catalyst types such as transition metal phosphides (TMP) are proved to be prominent competitors in this domain showing versatile advantages. Despite significant advancements, investigating new catalytic systems for oil hydrotreatment continues to be a challenging field of study.

2. Hydrotreatment Process

Hydrotreatment has been investigated as a potential substitute method for producing biodiesel. Using this process, triglycerides and fatty acids are converted into intermediate distillates—straight-chain alkanes that fall between n-C₁₅ and n-C₁₈ and are appropriate for use in diesel fuel applications. These resultant alkanes have a high cetane number exceeding 98, according to studies [19]. These processes can be broadly divided into two categories: co-processing of vegetable oil with refinery fractions generated from crude oil [20,21,22], and hydrotreating of pure vegetable oil [20,23,24,25]. Hydrotreatment is a broad term that includes a variety of processes, such as the decarboxylation (DCO₂) and decarbonylation (DCO) of fatty acids, as well as the hydrogenation and hydrodeoxygenation (HDO) of triglycerides. The hydrotreatment products had three phases: vapour, aqueous, and organic oil, based on these chemical interactions. The organic oil phase comprised the HVO product of hexadecane (n-C₁₆H₃₄) and octadecane (n-C₁₈H₃₈), while the vapour phase contained the co-products of carbon monoxide (CO), carbon dioxide (CO₂), and propane (C3H8). Moreover, water makes up the aqueous phase (H₂O).

Propane and intermediate free fatty acids are formed when the double bonds in fatty esters are hydrogenated and the molecules disintegrate, as is typically the case in the first reaction inside the hydrotreatment reactor [26]. Following this, there are three processes by which oxygen is eliminated: hydrodeoxygenation, decarboxylation, and decarbonylation [27]. While the other two processes produce alkanes with an odd number of carbon atoms together with CO₂ and CO, hydrodeoxygenation produces water and alkanes with an even number of carbon atoms [28].

On top of that, the production of HVO via hydrotreatment demands a significant quantity of hydrogen, typically sourced from steam reforming of hydrocarbons. This circumstance underscores the importance of exploring alternative methods that can effectively generate hydrogen, utilising renewable energy sources. A widely studied approach involves utilising low-carbon electricity to facilitate the production of synthetic hydrocarbons. This method has been extensively investigated in conjunction with both gaseous and liquid fuels [16,17,18,19].

Currently, the available literature primarily focuses on pathways that integrate electrolysis with fuel production, emphasising hydrocarbon synthesis through methanation [23], Fischer–Tropsch processes [20], and methanol/dimethyl ether (DME) synthesis [24,25]. Given that numerous studies indicate the potential of HVO production to significantly contribute to decarbonising the transportation sector [21], with forecasts suggesting the gradual replacement of conventional biodiesel with HVO [22,26], there is a pressing need for a comprehensive examination of the techno-economic and environmental performance of this process.

Given that the hydrogen cannot be obtained from these kinds of excellent techniques, such as electrolysis, because of their expensive and complicated structural requirements, high-selective catalysts will be required in order to convert vegetable oil triglycerides into high-yield, value-added hydrocarbons that require less hydrogen to produce than under initial conditions. Therefore, it is necessary to give a clear representation of the specifics of each phase in the hydrotreatment process in order to be able to use these catalyst types efficiently.

Firstly, the conversion of triglycerides over hydrotreating catalysts, in the presence of hydrogen, involves intricate reaction pathways consisting of parallel and/or consecutive steps such as saturation, cracking, decarboxylation, decarbonylation, and/or hydrodeoxygenation as shown in Figure 1. During the first phase, the double bonds present in the triglycerides underwent saturation with hydrogen. Fatty acids containing double bonds in their chains, including palmitoleic acid (C16:1), oleic acid (C18:1), linoleic acid (C18:2), alpha-linoleic acid (C18:3) and eicosenoic acid (C20:1) were converted into palmitic acid (C16:0), stearic acid (C18:0), and arachidic acid (C20:0), respectively. After that, the hydrogenated triglyceride underwent degradation into various intermediates, including monoglycerides, diglycerides, and free fatty acids. Subsequently, these intermediates were converted into deoxygenated products [27]. The generation of n-alkanes from free fatty acids can occur through one or a combination of three distinct reaction pathways: decarboxylation, decarbonylation, and/or hydrodeoxygenation [28].

In the decarboxylation pathway, the carboxylic group in the free fatty acids is converted into straight-chain alkanes, releasing CO₂. Hydrogen is not required for the decarboxylation reaction. Conversely, in the decarbonylation pathway, alkanes are produced by the reaction of the carboxylic acid group in the free fatty acids with hydrogen, resulting in the formation of CO and water. Alkanes generated through decarboxylation and decarbonylation typically contain an odd number of carbons in their chains.

On the other hand, the hydrodeoxygenation route releases water and converts carboxylic acid with hydrogen to form alkanes with an even number of carbons. Therefore, the ratio of n-alkanes with odd carbon atom counts to n-alkanes with even carbon atom counts (e.g., n-C₁₇/n-C₁₈) can be used to determine the relative abundance of decarboxylation and decarbonylation processes as opposed to hydrodeoxygenation pathways. Isomerisation, cyclisation, and cracking are examples of secondary reactions that can take place in addition to the primary processes of decarboxylation, decarbonylation, and hydrodeoxygenation. These reactions can result in the generation of lighter hydrocarbons, aromatics, and iso-alkanes. Methanation and water–gas shift reactions can occur to light gaseous byproducts from hydroprocessing, such as carbon dioxide, carbon monoxide, and water. The molar ratio of CO₂/CO can be used to assess the methanation reaction, which is undesirable since it uses up important hydrogen, but it can also be used to determine which decarboxylation and decarbonylation routes are more prevalent.

Overall, the hydrotreated biodiesel production process necessitates a comprehensive approach with an improvement of each component to enhance its advantages and address its drawbacks, which will be discussed below. Ongoing intensive research and technological advancements aim to enhance process efficiency and economics.

However, the focus of this section, dedicated to hydrotreatment technology for biodiesel production, is primarily on scientific aspects rather than engineering considerations. The subsequent subchapters will address key factors crucial for achieving more efficient biodiesel production through hydrotreatment:

• The quality of feedstock, including parameters such as free fatty acid content and phosphatide content.
• Reaction conditions, encompassing factors like temperature, reaction time, and mixing intensity.
• Selection of the catalyst type. This latter aspect will be further elaborated, including a comparison of different catalytic systems, in Section 3.

2.1. Feedstocks

Moreover, these renewable HVO fuels, which utilise various vegetable oils as feedstock selection, have been categorised as second-generation biodiesel because they meet the requirements of EN 590 and ASTM D 975 [29]. HVO can be derived from abundant feedstock ranging from vegetable oils, and animal fats, to microalgae. Palm oil, one of the common feedstocks, has also been utilised for producing HVO, with a product yield exceeding 90%. Recently, many vegetable oils and fats, such as rapeseed and animal fats, have been explored for their technological and economic processes in producing hydrotreated renewable diesel. However, since these proposed biofuels primarily comprise edible food stocks, concerns regarding food security may arise due to potential disruptions in the food chain caused by biofuel production. The “food vs. fuel” dilemma, which pertains to the risk of compromising food supply due to the allocation of farmland for biofuel production, may further impact societal food security through changes in market-based incomes and food prices [30].

Therefore, alternative inedible oils have garnered greater attention as aviation fuel substitutes for fossil-based diesel due to their flexibility and competitiveness with food production [12]. Despite significant progress in potential feedstocks of inedible oils, such as sludge palm oil (SPO), it has not yet reached a mature stage. The global production of palm oil has continued to increase. In 2020, Malaysia was the world’s second biggest producer of CPO, with an output of 19.14 million tons. The by-product of palm oil production, palm oil mill effluent (POME), generates approximately 2.5–3.0 tons per ton of CPO production. The residual oil separated during the initial stage of the POME discharge pond, known as SPO, is abundant and accounts for approximately 2% of total palm oil production, with a potential annual yield of 41 million tons [31]. As a low-cost potential substrate for biodiesel production, SPO has a high free fatty acid (FFA) content of more than 40% [32]. Although numerous studies have focused on the techno-economic analysis of biodiesel and other bio-based diesel, such as bioethanol, limited research has been conducted on hydrotreated vegetable oil (HVO). Thus, utilising SPO as a feedstock holds promise for sustainability through waste utilisation and lower production costs.

SPO refers to the residual oil floating upon the initial discharge of POME during CPO production, characterised by a high free fatty acids (FFAs) content [33]. The fatty acid composition of SPO is sourced from Nasaruddin et al., with chemical components such as palmitic acid, stearic acid, oleic acid, and linoleic acid. Additionally, triolein (C₅₇H₁₀₄O₆), a major triglyceride in palm oil, constitutes 48% of SPO content.

Cui Jun et al. illustrate the base and best cases with the usage of the non-edible industrial waste of SPO, in which it required 2% and 1.4% higher oil feedstock than the process of utilisation of refined bleached deodorised palm oil (RBDPO) as feedstocks, respectively, which showed the same desired HVO production rate of 25,000 kg/h [34]. However, in both cases, the hydrogen requirement was 10.5% and 29.2% lower, respectively. In the best case, the recovery of hydrogen can decrease by 21% of the raw hydrogen feed, and with the simulated results, 96.3% and 97.4% of CO and CO₂ emissions can be reduced by this SPO-HVO production process.

Additionally, one of the primary sources utilised for producing HVO is waste vegetable oil (WVO). In Europe, the estimated annual production of WVO ranges between 700 and 1000 kilotons [35]. The main challenges associated with utilising these resources revolve around their collection and treatment. Converting WVO into biofuels, especially biodiesel, facilitates the implementation of the circular economy concept and reduces reliance on virgin oils. Since WVO is classified as a residue, its utilisation for biodiesel production could significantly contribute to meeting the Renewable Energy Directive (RED) targets for integrating renewable sources into the transportation sector. Notably, biofuels derived from residues benefit from double counting in terms of their energy content, making them neutral in greenhouse gas emissions [36,37].

Feedstock pre-treatment typically involves three main processes: neuralisation, degumming, and bleaching [38,39]. The presence and quantity of impurities in the waste oil affect the pre-treatment procedure. Free fatty acids, which can act as contaminants, are commonly removed through a neutralisation step employing alkaline solutions [40]. Other impurities, such as calcium and magnesium salts of phosphatic acid, are often nonhydratable and thus challenging to eliminate. Degumming is the process used to remove all types of phosphatides, including nonhydratable ones. In cases where the phosphatide content is low (<20 ppm), such as in palm oils, dry degumming processes involving concentrated phosphoric acid can be employed. However, if the phosphatide content is higher, alternative methods like acid or EDTA degumming may be necessary. The final stage, bleaching, aims to reduce the presence of non-converted phosphatides, trace metals, and other contaminants in the oil [40,41].

Additionally, the physico-chemical characteristics and detailed composition of the feedstock are crucial for determining the final product’s quality (as shown in

Table 1. The physico-chemical characteristics of the vegetable oil feedstock.

Property	Value Range	Importance
Density (g/cm3 at 15 °C)	0.88–0.93	Influences fuel atomisation and combustion efficiency
Viscosity (mm2/s at 40 °C)	30–50	Affects fuel flow and injection system performance
Acid Value (mg KOH/g)	<2	High values result in catalyst poisoning
Iodine Value (g I2/100 g)	80–140	Illustrates the unsaturation degree, affecting oxidative stability
Saponification Value (mg KOH/g)	180–200	In relation to the average molecular weight of the fatty acids
Moisture Content (%)	<0.5	High moisture can lead to hydrolysis and catalyst deactivation
Phosphorus Content (ppm)	<10	High phosphorus can cause catalyst poisoning
Metal Content (ppm)	<5 (Na, K, Ca, Mg)	Metals can form deposits and poison the catalyst

and

Table 2. The composition of the vegetable oil feedstocks.

Component	Sludge Palm Oil (%)	Waste Vegetable Oil (%)
Palmitic acid (C16:0)	25–30	10–20
Stearic acid (C18:0)	3–5	4–6
Oleic acid (C18:1)	35–45	30–50
Linoleic acid (C18:2)	5–10	10–30
Linolenic acid (C18:3)	<1	0–3
Free Fatty Acids (FFA)	2–10	5–15

). The tables below provide instructive data on the various properties of vegetable oils and their compositions [42].

Alternatively to the traditional vegetable oils, as discussed earlier, palm oil derivatives such as SPO proved themselves as renewable feedstock for the hydrotreatment process and waste vegetable oil feedstock has gained some attention too [43].

From the last table, we can see that SPO has high saturated fatty acid content (palmitic acid), which ensures good oxidative stability and high cetane number in biodiesel; meanwhile, WVO feedstock poses its oleic and linoleic acid contents, which can produce high-quality biodiesel if properly pretreated. Overall, the representation of available and relevant vegetable oil feedstock and their characteristics are provided for the catalytic hydrotreatment process in this section.

2.2. Influence of Reaction Time, Pressure, and Temperature

It is essential to consider several key process variables when producing biodiesel from vegetable oils, including the reaction time, pressure, and temperature. Among these variables, pressure drops of hydrogen during the hydrotreatment process play a crucial role in determining biodiesel yield. Bambang et al. conducted the hydrotreatment of soybean oil with a clear representation of pressure and temperature changes with reaction time in the presence of 57.6 wt% Ni/SiO₂–Al₂O₃. The hydroprocessing trials were carried out utilising a specially constructed, high-pressure batch reactor setup. This reactor, which was cylindrical, had dimensions of 34.5 mm in diameter and 117 mm in height, resulting in a total volume of 109 cm³. Before each trial, 28.1 g of soybean oil was loaded into the oil supply tank, followed by a purging process with N₂ for a minimum of 30 min to eliminate any dissolved oxygen from the oil and the feed tank itself. Afterwards, the reactor was pressurised with H₂ to the desired pressure for the experiment, which was 9.2 MPa. Once this target pressure was attained, the contents of the reactor were stirred using the magnetically driven stirrer, while simultaneously raising the temperature of the reactor to the desired level for the experiment, which was 400 °C. This temperature was maintained for one hour. Subsequently, once the reaction temperature reached 400 °C, the reaction proceeded for another hour [27].

During the process, a significant hydrogen pressure drop at around 5.3 MPa was observed in the temperature range of 100–130 °C, as shown in Figure 2. And after, a second drop in hydrogen pressure at around 3.5 MPa occurred during the rise in temperature, in the range of 270–330 °C. These phases correspond to the former saturation of triglycerides with hydrogen and the latter degradation of hydrogenated triglycerides into free fatty acids, respectively. Consequently, with increasing temperature and pressure in a sufficiently lasting time span of close to 2 h, the atoms of hydrogen start to collide and react more with double bonds of triglycerides which favours the hydrogenation step of the hydrotreatment process.

However, when soybean oil underwent hydrotreatment using either NiMo or Pd catalysts, the distillation characteristics exhibited a flat distribution across a wide range of recovery within a limited range of boiling points of 290–330 °C. The fractions with higher boiling points ranging from 380 to 450 °C might represent partially reacted intermediates between triglycerides and alkanes. The similarity observed in the final boiling points of the hydroprocessed products to that of soybean oil could be attributed to the presence of unreacted triglycerides. Another potential explanation for the heavy fraction in the liquid product could be linked to the oligomerisation and aromatisation of reaction intermediates containing double bonds in their molecular structures [44].

Songphon et al. reported that reducing the initial H₂ pressure from 40 to 20 bar has minimal impact on the conversion and liquid product for the NiS catalyst [45]. This could be attributed to the low hydrodeoxygenation activity of NiS, which relies heavily on H₂ pressure, as indicated by C-O cleavage [46]. Conversely, for the NiP catalyst, a decrease in conversion from 77.4% to 72.5% was observed alongside a notable increase in the C_n−1/C_n ratio from 1.6 to 3.4 demonstrating a strong dependency on initial H2 pressure. NiP, being the catalyst showing the most promising hydrodeoxygenation activity among the three, experiences a decline in the given activity with decreasing H₂ pressure due to reduced hydrogen availability on its surface.

The impact of pressure is more pronounced with the NiC catalyst, where conversion drastically drops from 78.9% to 65.4%. This decrease is attributed to significant H₂ consumption through methanation reaction, potentially leaving insufficient H₂ for deoxygenation and methanation concurrently.

Regarding gas product distribution, an increase in H₂ pressure results in higher CH₄ contents across all three catalysts, indicating the favourability of the methanation reaction under elevated H₂ pressure. The CO₂ fraction increases with decreasing H₂ pressure only for NiP and NiC catalysts. Although the explanation for this CO₂ fraction is not straightforward, one plausible explanation could be the consumption of H₂. As H₂ is consumed, the limitation of H₂ forces the reaction towards CO₂ production [45].

Overall, the effects of initial hydrogen pressure on the performance of different catalysts in hydrodeoxygenation reactions vary drastically for each one of them. While reducing H₂ pressure has minimal impact on conversion and liquid product for one catalyst, it significantly affects conversion and product distribution for the other. Consequently, we can observe the general logic behind the hydrotreatment process when higher H₂ pressure favours methanation reactions, while lower pressure increases CO₂ production, possibly due to H₂ consumption. The use of various catalysts is presented to provide a comprehensive view of process conditions, which can follow one pathway for one system and may change for another.

3. Choice of Catalyst for Hydrotreatment

The catalytic HVO is typically conducted within a temperature range of 300 °C to 450 °C, while the applied hydrogen pressure falls between 3 MPa and 20 MPa [47,48]. Various noble or transition metal catalysts such as Pd, Pt, Co-Mo, and Ni-Mo can be employed for this process, and these catalysts, along with temperature and pressure conditions, significantly influence the distribution of products [49].

In a continuous hydrotreatment process, the vegetable oils undergo three main steps. Initially, the substrate is mixed and pre-saturated with hydrogen under high temperatures [50]. Subsequently, the gas–liquid mixture is introduced into the reactor, where it reacts over a fixed bed of catalyst in a trickle bed configuration. Finally, the resulting products are directed to the separator after cooling, allowing for the removal of light by-products and unreacted hydrogen dissolved in the liquid phase. Sulphided Co-Mo, Ni-Mo, and Ni-W catalysts supported on γ-Al₂O₃ typically exhibit higher activity compared to monometallic catalysts in a continuous process manner [49]. Among these, Co-Mo and Ni-Mo catalysts tend to promote the hydrodeoxygenation reaction pathway. Specifically, in a hydrotreating reaction involving soybean oil, it was observed that using Co-Mo/γ-Al₂O₃ resulted in one of the lowest C₁₇-to-C₁₈ product ratios among the results obtained with other mono- and bi-metallic catalysts [27].

Additionally, researchers are enhancing the quality of bio-oils by transforming them into more stable hydrocarbons, This process involves using specific catalysts for reactions such as hydrodeoxygenation (HDO), hydrocarbonylation, or hydrodecarboxylation [47,51,52]. Thus, it requires heterogeneous catalysts as mentioned above along with high temperature and pressure, in order to remove oxygen from vegetable oil and produce carbon atoms in the C₁₅–C₁₈. Combining vegetable oil with heavy and gas oil during processing leads to the elimination of nitrogen and sulphur from the mixture [53,54,55,56,57,58,59]. Also, studies have investigated the impact of factors like liquid hourly space velocity, hydrogen-to-oil ratio, and pressure on the production of green diesel from waste cooking oil using an Ni-Mo catalyst, which showed high conversion and diesel selectivity as efficient hydrotreating catalysts [60].

3.1. Heterogeneous Catalysts

Traditional heterogeneous hydroprocessing catalysts like NiMo and CoMo have long been utilised for hydrodesulphurisation (HDS) in removing sulphur from petroleum products. Interestingly, these same catalysts have shown effectiveness in the hydrogenation and deoxygenation of triglycerides. However, catalyst deactivation poses a significant challenge, particularly due to the high oxygen content in triglycerides, which can lead to sulphur leaching from the catalyst surface [61]. To maintain catalyst activity, sulphiding agents such as H₂S and dimethyl disulphide are often introduced into the reaction system [62,63,64,65]. However, the optimal amount of sulphiding agent must be carefully determined to achieve the desired catalyst activity and product quality. Moreover, these compounds are highly toxic and can pose significant health risks, particularly to the nervous system. Therefore, non-sulphided heterogeneous catalysts such as nitride catalysts, Pd and Pt are gaining interest in biofuel production as they eliminate the need for sulphiding agents when processing biomass-derived feedstocks, thus preventing the deactivation of sulphided metal catalysts [66,67,68]. Previously, various metal catalysts including Pd, Ni, Ru, Ir, Os, and Rh supported on alumina, silica, and activated carbon, as well as some alloys and bimetallic catalysts, have been investigated for stearic acid deoxygenation [69].

On top of that, the conditions under which reactions occur and the types of catalysts employed have significant impacts on the makeup and quality of the resulting liquid product. Gusmao et al. conducted research on the hydrocracking of soybean and babassu oils to produce hydrocarbons using diminished Ni/SiO₂ and sulphided NiMo/γ-Al₂O₃ catalysts within a batch reactor, operating within temperatures ranging from 350 to 400 °C and hydrogen pressures between 1 and 20 MPa. The primary products of the reaction were aliphatic hydrocarbons generated through total decarbonylation, decarboxylation, or hydrogenation processes [70]. Da Rocha Filho et al. similarly investigated the hydrocracking of soybean oil along with other vegetable oils such as maracuja, tucuma, buriti, and babassu oils, utilising sulphided NiMo/γ-Al₂O₃ in a batch reactor [71]. The resulting products mainly comprised n-alkanes (66–67 wt%), cycloalkanes (up to 13 wt%) and alkyl aromatics (up to 4 wt%) after a 2 h reaction at 360 °C and an initial hydrogen pressure of 14 MPa. Huber et al. studied the hydrotreating of sunflower oil using a flow-type reactor with a sulphide NiMo/Al₂O₃ catalyst, operating at temperatures ranging from 300 to 450 °C and a hydrogen pressure of 5 MPa [28]. Under optimised conditions, the molar yield of carbons from n-C₁₅ to n-C₁₇ was 71%. Simacek et al. investigated the hydroprocessing of rapeseed oil using a flow-type reactor at temperatures of 260–340 °C and a 7 MPa hydrogen pressure, the liquid product consisted of more than 70% by weight of n-C₁₇ and n-C₁₅, alongside minor constituents such as n-alkanes with carbon numbers lower than 15 (n-C₁₅), iso-alkanes of C₁₆–C₁₈, and cycloalkanes [72].

3.2. Nickel-Based Catalysts

Nickel–molybdenum sulphides are traditional catalysts for hydrotreating that have undergone extensive research and industrialisation. Nevertheless, sulphiding agents pose significant risks to living organisms and contribute to environmental issues. Additionally, they have the potential to react with reactants and intermediates, leading to the formation of undesirable sulphur-containing products [73,74]. Hence, there is a current exploration into catalysts that do not require the addition of sulphiding agents [51,75,76]. So newer options such as nickel phosphide (NiP) and nickel carbide (NiC) have emerged for desulphurisation purposes [77,78].

Recent studies have focused on NiP for removing oxygen from various model compounds such as dibenzofuran, methyl laurate, and methyl oleate [79,80,81]. NiP, particularly, has shown promise in mentioned deoxygenation reactions with reduced levels of methanation and cracking compared to metallic nickel, thanks to its distinctive characteristics as the ligand effect and ensemble on metal sites [82]. NiP exhibits both metallic and acidic properties, potentially resulting in a synergistic effect. Transition metal phosphide, especially molybdenum phosphide (MoP), has demonstrated high conversion rates in hydrodeoxygenation of methyl laurate between others like Ni, Co, Fe, Mo, and W phosphides. Although NiP may exhibit slightly lower conversion rates (ca. 72%) compared to MoP (ca. 85%), it consumes less hydrogen and favours specific deoxygenation pathways such as decarboxylation/decarbonylation (C_n−1/C_n = 4.17), while the most prioritised reaction pathway for MoP (C_n−1/C_n = 0.04) was hydrodeoxygenation as well [83].

Recently, Zhang et al. prepared NiP supported on γ-Al₂O₃, claiming advantages such as a large specific area, good thermal stability, intrinsic acidity, and suitability for hydroprocessing [82]. High conversion rates with 95% methyl laurate with considerable selectivity to C₁₁–C₁₂ can be achieved with 10% Ni loading. Research on NiC as a catalyst for the hydroprocessing of triglycerides/fatty acids or their derivatives is limited. Some transition metal carbides like Mo₂C and W₂C have shown promise for hydrodeoxygenation of methyl stearate and stearic acid compounds, respectively, by performing great hydrotreating activity and having the advantage of an inexpensive catalyst over noble metals [84,85]. However, Manoli et al. discovered that NiC/Al₂O₃ (82% initial conversion) exhibited better hydrodesulphurisation activity compared to Mo₂C/Al₂O₃ (initial conversion of 3%) [78]. NiC is anticipated to demonstrate high catalytic activity for deoxygenation, including a direct comparison with NiS and NiP under different operating conditions.

Songphon et al. provided the XRD patterns of NiS, NiP, and NiC supported on γ-Al₂O₃ catalysts as illustrated in Figure 3. In the final sample, the NiC catalyst showed its cubic structure according to JCPDS card 14–0020. In characteristic peaks at 2θ = 44.5°, 51.9° and 75.5°, there was a representation of NiC (111), (200) and (311), respectively. Combined forms of Ni₃S₂ and Ni₃S₄ were shown in peak patterns of NiS catalysts. In the case of the NiP catalyst, the XRD pattern represents the mixed forms of Ni₁₂P₅ and Ni₃P. The overlap with peaks of γ-Al₂O₃ support reflects the phosphide phases [45].

In

Table 3. Characteristics of the catalysts and aluminium oxide support. Reproduced with permission, Ref. [45].

Catalyst	Surface Area (m2 g−1)	Mean Pore Diameter (nm)	Total Pore Volume (cm3 g−1)
NiS/γ-Al2O3	75.9	5.9	0.11
NiP/γ-Al2O3	54.7	3.3	0.05
NiC/γ-Al2O3	75.8	5.6	0.07
NiO/γ-Al2O3	143.3	8.2	0.35
γ-Al2O3	179.0	8.8	0.40

, there is the specific surface area, mean pore diameter, and pore volume of hydrotreatment catalysts with NiO/γ-Al₂O₃ and bare γ-Al₂O₃. These values decreased during the loading of NiO into γ-Al₂O₃. Consequently, the specific surface area, mean pore diameter, and pore volume drastically declined after the catalyst was activated, for example, the surface area of NiC and NiS was 76 m²/g while 55 m²/g for NiP. The changes in all these values following the activation of NiO into various forms indicate a transformation in the active phase, which was not clearly detected by XRD, as seen in the case of NiP (Figure 3) [45].

Overall, a series of nickel catalysts, namely nickel sulphide (NiS), nickel phosphide (NiP), and nickel carbide (NiP), were examined for the hydrotreating of spent coffee oil to produce bio-hydrotreated fuel (BHF). Catalytic tests were conducted at temperatures ranging from 375 to 425 °C and initial H₂ pressures of 20 to 40 bar, with reaction times of 0 to 3 h. The catalytic activity followed the order NiC > NiP > NiS, with NiC showing a tendency to promote cracking reactions, resulting in higher yields of gasoline and gaseous products. Conversely, NiS, while exhibiting the lowest oil conversion, favoured diesel yield with lower methanation and cracking activity [45].

Decarbonylation was found to be the major route for deoxygenation of coffee oil across all catalysts, surpassing decarboxylation and hydrodeoxygenation. The ratio of (DCO + DCO₂) to HDO, represented by C_n−1/C_n, decreased in the order NiS > NiC > NiP. Intermediate ketones (approximately 3 wt%) were detected with NiP, likely formed through alcohol rearrangement and keto-enol tautomerism.

NiS-catalysed liquid products contained a significant amount of aromatics (4 wt%) along with isomerisation products (0.9 wt%), whereas NiP and NiC catalysts showed only trace amounts of these compounds. Physiochemical analysis of the diesel fraction revealed properties consistent with commercial bio-hydrogenated diesel according to Neste MY Renewable Diesel (NExBTL) standards, with satisfactory density and kinematic viscosity conforming to specifications. Given that the main products are straight-chain hydrocarbons, a high cetane index (>110) could be attained [45].

3.3. Palladium-Based Catalysts

In the past, numerous metal catalysts like palladium (Pd), nickel (Ni), ruthenium (Ru), iridium (Ir), osmium (Os), and rhodium (Rh), supported on materials such as alumina, silica, or activated carbon, as well as certain alloys and bimetallic catalysts, have been investigated for stearic acid deoxygenation. Among these catalysts, it has been observed that Pd/C exhibits the highest activity at 300 °C under 17 bar of helium in a semi-batch reactor, achieving a selectivity of 95% towards heptadecane with complete conversion. A variety of oxygenated compounds have been utilised as starting materials for hydroprocessing, including fatty acids [55,86,87,88] and their esters or direct triglycerides [70,71]. Triglycerides from vegetable oils can be hydrogenated and decomposed into various intermediates (such as monoglycerides, diglycerides, and carboxylic acids), with the glycerol backbone of triglycerides being transformed into propane.

Traditionally, inedible oil and waste cooking oil are primarily utilised for biodiesel production to mitigate the food vs. fuel debate. However, palm oil offers the advantage of efficient planting, harvesting, and collection processes. Worapon et al. conducted research on optimal operating conditions for the hydroprocessing of various palm oil feedstocks, namely crude palm oil (CPO), degummed crude palm oil (DPO), and palm fatty acid distillate (PFAD), using both commercial 5 wt% Pd/C and synthesised NiMo/γ-Al₂O₃ catalysts were examined [89]. In further elaborations, the NiMo/γ-Al₂O₃ catalyst is used as a benchmark for the purpose of comparing with the palladium catalyst and defining the main advantages of the focused catalyst.

Table 4. Characteristics of catalysts. Reproduced with permission [92].

Catalyst	Surface Area (m2 g−1)	Mean Pore Diameter (nm)	Crystallite Size (nm)	Total Pore Volume (cm3 g−1)	Average Particle Size (μm)
Pd/C	820	4.2	3.6	0.85	15
NiMo/γ-Al2O3	160	7.1	1.3	0.22	17

provides details on the specific surface area, pore volume, mean pore diameter, and crystallite size of the catalysts. The XRD patterns of Pd/C and unsulphided NiMo/γ-Al₂O₃ catalysts are depicted in Figure 4. The characteristic peaks at 2θ = 40.1°, 46.7° and 68.1° correspond to the (111), (200), and (220) planes of Pd, respectively. The crystallite size of Pd, calculated from the broadening of the Pd (111) peak using the Debye–Scherrer formula, is 3.62 nm, as indicated in Table 4. Figure 4 illustrates the XRD patterns of the unsulphided NiMo/γ-Al₂O₃ catalyst as a comparison, revealing the absence of distinct metal phase characteristic peaks. This suggests that nickel oxide and molybdenum oxide were highly dispersed on the catalyst surface. Despite higher % loading of Ni and Mo, a similar XRD pattern was observed, consistent with findings by Wang et al., who used comparable metal precursors [90]. However, it is worth noting that NiO and Ni₂O₃, as the most common Ni oxides, may also exist in microcrystalline or amorphous phases, while NiMo/γ-Al₂O₃ peaks may overlap with peaks of γ-Al₂O₃ and would be observed only at high loading [91].

The performance of hydroprocessing over Pd/C was compared with another active phase, non-sulphided NiMo/γ-Al₂O₃. Degummed crude palm oil and palm fatty acid distillate were selected as feedstocks for the investigation. For both feedstocks, the optimal operating condition was found to be the same: 375 °C, 40 bar for 1 h. The maximum diesel yields of 70% and 76% were obtained for DPO and PFAD, respectively. Indeed, the pathways of deoxygenation or decarboxylation/decarbonylation were strongly influenced by the nature of the catalyst, operating temperature, and reaction pressure [93]. The Pd/C catalyst predominantly produced decarboxylation hydrocarbon products with increasing temperature [94,95,96], while NiMo/γ-Al₂O₃ primarily yielded hydrodeoxygenation products [97].

However, when comparing the catalysts using PFAD as feedstock, Pd/C demonstrates higher catalytic activity, indicated by a higher diesel yield, compared to NiMo/γ-Al₂O₃. This finding aligns with that of Snare in their study on the deoxygenation of stearic acid over NiMo/γ-Al₂O₃, Pd/C and other catalysts [69]. Low conversion (8.6%) and C₁₇ selectivity (23%) were achieved using NiMo/γ-Al₂O₃ as the catalyst, whereas Pd/C showed high activity (100% conversion) with 99% C₁₇ selectivity. Higher catalytic activity of Pd/C over NiMo/γ-Al₂O₃ was also reported for the deoxygenation of carpylic acid. Therefore, it can be concluded that Pd/C is suitable for fatty acid feedstock.

On the contrary, in the case of DPO triglyceride feedstock, NiMo/γ-Al₂O₃ exhibits higher catalytic activity than Pd/C catalyst. Some contradictions in the rate-determining step in the literature were reported. The study of hydroprocessing of tricaprylin and carpylic acid over NiMo/γ-Al₂O₃ suggested that the overall rate of the reaction was governed by the deoxygenation (decarboxylation/decarbonylation or hydrodeoxygenation) step [98]. Meanwhile, the study by Madsen et al. suggested that the hydrodeoxygenation of tripalmitin catalysed by Pt/γ-Al₂O₃ was limited by the formation of a palmitic acid intermediate [99].

On top of that, Domínguez-Barroso et al. performed research on the production of hydrocarbons from sunflower oil under subcritical water conditions and without the need for extra hydrogen input, the Pt-Ni/Al₂O₃ catalyst, typically employed in steam reforming, is utilised alongside the Pd/C catalyst, typically used in hydrogenation [100]. From that, the concurrent use of Pt-Ni/Al₂O₃ and Pd/C catalysts in a dual fixed-bed setup facilitated additional decarboxylation of free fatty acids (FFA), resulting in the production of a C17-paraffinic hydrocarbon with a cetane number of 48.8, a density of 0.774 g cm⁻³ and a higher heating value of 47.53 MJ kg⁻¹. These characteristics are similar to those of commercial HVO.

Given that decarboxylation/decarbonylation requires less hydrogen compared to hydrodeoxygenation, a homemade Pd/C catalyst (1% w/w Pd, structured as a spiral-wound carbon catalyst) was incorporated alongside the Pt/Ni-Al₂O₃ catalyst to enhance the deoxygenation through decarboxylation of the semi-solid product. It is important to note that the deoxygenation extension for sunflower oil was notably higher at 300 °C after 4 h of reaction under these conditions. An increase in pressure (103 bar) was observed due to the rise in the gas-phase reaction, although the distribution of gaseous products remained similar to that observed over the Pt/Ni-Al₂O₃ catalyst, albeit with reduced methane and ethylene percentages. The combined catalytic system showed improvements in hydrogenation and deoxygenation reactions, achieving a 71% conversion for decarboxylation and an 86% overall conversion of vegetable oil into the desired product [100].

Operating under subcritical water conditions and hydrogen limitation, a proposed scheme for the transformation of vegetable oil using the combined catalytic system, Pt/Ni-Al₂O₃ + Pd/C, is illustrated in Figure 5. The process begins with non-catalytic hydrolysis of triglycerides from sunflower oil (step 1), driven by equilibrium in the liquid phase, leading to the breakdown of triglycerides in the presence of water, yielding glycerol and three fatty acid molecules. Sequential steps involve hydrogen production through glycerol aqueous phase reforming (APR) catalyst by Pt/Ni-Al₂O₃ (step 2), followed by gas-phase water gas shift (WGS) and methanation reactions. Glycerol serves as a hydrogen donor for in situ hydrogenation of double bonds (step 3) and for deoxygenation via decarboxylation (step 4), shortening the carbon chain to form the derived paraffinic hydrocarbon [100].

The combination of Pt/Ni-Al₂O₃ + Pd/C enhances in situ hydrogen consumption, avoiding side reactions such as methanation, light hydrocarbon formation, and carbon monoxide (CO) production in the gas phase, as well as water gas shift. This approach enables the establishment of the process sequence involved in triglyceride hydroconversion without external hydrogen supply, resulting in a product with properties similar to commercial HVO and suitable for use as a biofuel. Basically, thanks to this work, it can be seen that the combination of palladium with platinum catalyst provided an illustration of the former catalyst’s benefits in relation to the latter.

Additionally, Songphon et al. conducted research on the hydrotreatment of the oil extracted from spent coffee grounds over two catalysts Pd/C and NiMo/γ-Al₂O₃ with different operating parameters [92]. Although increasing the H₂/oil ratio promoted the overall reaction and hydrodeoxygenation activity (as indicated by a decrease in C_n−1/C_n ratio) for both catalysts, hydrocracking was notably enhanced over Pd/C, resulting in a significant increase in gasoline yield. Moreover, Pd/C yielded a higher olefin content in the liquid product (22.3 wt%) compared to NiMo/γ-Al₂O₃ (4.8 wt%). However, NiMo/γ-Al₂O₃ exhibited higher isomerisation activity. The amount of isoparaffins catalysed by NiMo/γ-Al₂O₃ and Pd/C were 10.8 and 1.7 wt%, respectively.

3.4. Cobalt-Based Catalysts

Sebos et al. assessed the effectiveness of a commercial HDS (hydrodesulphurisation) CoMo/Al₂O₃ catalyst in hydrotreating cottonseed oil in the presence of desulphurised diesel to improve the cetane number of the resulting diesel fuel which in turn will not react with the catalyst system. The experiments were carried out using a small-scale trickle bed reactor, with a focus on monitoring the conversion of the oil. Additionally, the study accounted for catalyst deactivation and incorporated it into the analysis of reaction kinetics [101].

The liquid feed comprised 10 wt% refined cottonseed oil in desulphurised diesel (S < 50 ppm). The catalyst employed was a commercial CoMo/Al₂O₃ in cylindrical form with an average diameter of 1.4 mm, primarily recommended by suppliers for producing ultra-deep desulphurised diesel oil. The catalytic bed, diluted, spanned 6.1 cm in length and included 5 g of catalyst and 36.2 g of inert carborundum (SiC) particles with an average diameter of 0.25 mm. Glass spheres, measuring 3 mm and 4 mm in diameter, were positioned before and after the catalyst bed to ensure uniform liquid distribution and provide support. Loading of the bed was compact as a catalyst and diluent carborundum was added sequentially in the continuously agitated reactor.

Prior to use, the catalyst was dehumidified under hydrogen at a flow rate of 15 NL/h for 2 h at 100 °C. The catalyst was then sulphided in situ by ramping the temperature to 340 °C at a rate of 20 °C/h, with a gas flow of 15 NL/h and liquid flow of 20 g/h diesel. Presulphidation continued for 4 h at 340 °C. Following activation, experiments were conducted over 43 days, with each lasting 10–12 h daily. Deactivation was monitored by repeating standard experiments. At the onset of each experiment, high liquid and gas flow rates were employed to saturate the catalyst bed and prevent liquid unevenness during operation. Steady-state conditions were typically reached after 4–6 h of continuous operation, depending on liquid and gas flow rates. Between daily experimental intervals, the reactor was gradually cooled under hydrogen flow and maintained at a pressure of 30 bar overnight at room temperature [101].

To investigate the kinetics of the hydrodeoxygenation (HDO) process converting triglycerides from the feedstock into diesel-like hydrocarbons under the conditions of the experiments, he assumes that plug flow is a suitable approximation within the reactor. This is because the dilution of the catalyst bed with fine particles (SiC) guarantees complete wetting of the catalyst and prevents liquid unevenness or axial dispersion. Catalyst deactivation is tracked by computing the activity coefficient for each recurrent standard experiment. The parameters for the standard experiments were set as follows: reactor temperature at 320 °C, total reactor pressure maintained at 30 bar, H₂S partial pressure at 1 vol%, mass feed rate set to 40 g/h, and H₂ gas flow rate ranging between 23 and 24 NI/h.

$$\propto =\frac{k_{{HDO}_i}}{k_{{HDO}_o}}$$

Here ∝ represents the activity coefficient, with subscripts o and i denoting the reference and current values, respectively.

Figure 6 illustrates the activity of the catalyst and the conversion of ester over time. The catalyst underwent stabilisation during the initial 100 h, after which its activity was monitored as an activity coefficient value set as the reference one (∝ = 1). As depicted in Figure 6, the catalyst’s activity gradually decreases up to 300 h of operation, experiencing a reduction of approximately 13% compared to the reference value. Subsequently, it proves that the performance of the catalyst was rather stable to reach a steady state beyond this period.

In addition to that, Krause et al. investigated the hydrodeoxygenation reaction using model compounds that mimic aliphatic esters and acids found in wood-derived bio-oil [63,74,102,103]. In a similar vein, Laurent and Delmon conducted experiments focusing on the catalytic effectiveness of CoMo sulphide catalysts supported on various materials for removing oxygen-containing molecules [87,104,105,106,107].

3.5. Transition Metal Phosphides

Transition metal phosphides (TMP) represent a fascinating category of materials warranting exploration at the nanoscale due to their diverse range of properties and potential applications. When employed as catalysts, phosphides rich in metal exhibit outstanding activity in hydrogenation reactions. Several reviews have detailed the nature, structure, and synthesis of phosphides [108,109,110,111,112,113].

Historically, hydrotreatment reactions have relied on transition metal sulphides. However, these constituents require a source of sulphide and are not sufficiently stable in the presence of water. In recent times, phosphide catalysts and transition metal nitride, boride, and carbides have gained prominence [114,115]. Particularly since the late 1990s, transition metal phosphides (TMPs) have been recognised as highly efficient catalysts for hydrotreatment processes. These substances are composed of phosphorus and transition metals, amalgamating physical attributes such as the typical hardness and strength found in ceramics, with electronic properties resembling the metal conductivity characteristics. Additionally, transition metal phosphides are preferred as catalysts for HDO due to their numerous advantages over other active phases like transition metal sulphides and noble metals, notably in terms of cost-effectiveness and environmental friendliness [116]. Prior investigations involving CoP, Fe₂P, and Co₂P have demonstrated modest activity, while Ni₂P supported on silica exhibited high activity, especially in hydrogenation reactions [117,118]. WP and MoP, among other phosphides involving those with the metal iron, have been noted for their superior activity over metal sulphides [119].

Metal phosphide structures encompass compounds phosphorus alongside other elements from the periodic table, known as phosphides. The bonding characteristics within phosphides vary based on the counter-cation: ionic with earth metals, covalent or metallic with transition elements, and in the presence of main group elements are distinctly covalent. TMPs have been suggested as excellent catalysts due to their cost-effectiveness, abundance, and high efficiency [112,113]. In numerous fields, TMPs find utility as stable and effective catalytic materials.

Two primary families of TMPs have been identified based on their composition: metal-rich and phosphorus-rich phosphides. A plethora of TMPs have been characterised, including over a hundred monometallic TMPs spanning compositions from phosphorous-rich MP₁₅ to metal-rich M₄P, and around a hundred bimetallic metal phosphides as (M)_x(M’)_yP_z [109,120]. Concerning catalysis, the metal-rich phosphides are the most compelling compounds, possessing both ceramic and metallic properties, exhibiting adequate thermal and electrical conductivity, and demonstrating high chemical and thermal stability.

Initially, due to the comparable chemical and physical properties of TMPs, nitrides, and carbides, the same crystal structure is anticipated. However, the atomic radius difference between carbon and phosphorus or nitrogen leads to a distinct configuration, as phosphorus cannot occupy octahedral holes like its counterparts (C or N) can. Consequently, the crystalline configurations of phosphides feature metal atoms arranged in triangular prisms with phosphorus atoms, as depicted in Figure 7. In metal-rich phosphides, the structure shifts to a nine-fold tetrakaidecahedral (TKD) arrangement, with additional metal atoms positioned close to the centres of the prism’s vertical faces (Figure 7) [112,121].

From these foundational elements, various arrangements give rise to diverse structures observed in TMPs. Figure 8 provides an overview of the five primary TMP families. Among TMPs, the most extensively investigated compounds are those cantered around Co, Mo, Fe, and Ni. Nevertheless, recent studies have also delved into other TMPs, including W, Mn, Cu, and others.

Numerous catalysts have undergone examination for the upgrading of waste oils/fats and bio-oils, spanning from supported transition metals to phosphides, nitrides, and carbides. In spite of the diversity in hydrogenation phases, the most prominent catalysts commonly possess bifunctional properties, incorporating both acid and hydrogenation functionalities. This review focuses on examining phosphide catalysts presently used in HDO reactions.

Exploration of transition metal phosphides for HDO reactions is a relatively recent development, encompassing FeP [122], CoP [123], WP [122], Ni₂P [124], and Ru₂P [125]. The primary active phases of these HDO metal phosphides involve Bronsted sites and M+ (where M is a slightly positively charged transition metal) [124]. M+ serves as a Lewis acid site, participating in hydrogenolysis, demethylation, and hydrogenation reactions [126]. The Bronsted sites (PO-H) are formed thanks to phosphate precursors’ partial reduction, wherein active hydrogen forms are generated, albeit with lesser activity compared to those produced at metal sites. MoP exhibits bifunctional characteristics, displaying acidity and hydrogenation capabilities akin to noble metal deposits with acidic supports [127]. The proposed mechanism of exploitation of TMPs during HDO is outlined in Figure 9, supported by the literature findings [122,124]. Absorbed H₂ and oxy compounds became active on the M+ sites, and hydrogen atoms from both Bronsted PO-H sites and M+ sites reacted with the absorbed oxy intermediate variants yielding deoxygenated end products.

Oyama et al. investigated the HDO activity of the guaiacol model compound using various phosphides supported on silica [128]. The experiments were conducted under atmospheric pressure and in a fixed bed at 300 °C. The results indicated the following order of activity: Ni₂P > Co₂P > Fe₂P > WP > MoP. Considering the significantly lower cost compared to noble metals and their high activity, TMPs present an attractive option for bio-oil HDO processes. However, a drawback of phosphides is their susceptibility to oxidation by water (H₂O), which can lead to the formation of metal oxides or phosphates, subsequently causing catalyst deactivation.

In their study, Shi and his colleagues investigated methyl laurate during the hydrodeoxygenation reaction using various catalysts to produce C₁₂ and C₁₁ hydrocarbons [129]. These catalysts are nickel phosphides with TiO₂, HY zeolite, SAPO-11, SiO₂, Al₂O₃ and CeO₂ supports, all arranged with a ratio of 1.0 of Ni/P.

Figure 10A illustrates the modification of methyl laurate across various catalysts. With an increase in temperature from 300 to 340 °C, the conversion also increased for all catalysts. Notably, the highest conversion rate was observed in Ni₂P/SiO₂ at every temperature. The catalyst activity varied significantly with increasing temperature, indicating the influence of the supports on the activity of catalysts. Interestingly, a high Ni₂P/SiO₂ conversion rate attributed to the reduced SiO₂ support effect, likely due to its reducibility or lower acidity in relation to other supports.

Figure 10B presents the overall selectivity for C₁₂ and C₁₁ hydrocarbons across various catalysts. With increasing temperature, S(C₁₂ + C₁₁) also increased across all catalysts. Additionally, catalysts exhibiting higher conversion rates tended to demonstrate higher S(C₁₂ + C₁₁).

Figure 10C illustrates the influence of catalysts on the C₁₁/C₁₂ molar ratios. The HDO reaction of methyl laurate was observed in decarbonylation pathways. The molar ratio of C₁₁/C₁₂, reflecting the preference between these pathways, was strongly influenced by the catalysts. A higher C₁₁/C₁₂ molar ratio suggests dominance of the decarbonylation reaction. Furthermore, increasing the reaction temperature led to an increase in the C₁₁/C₁₂ molar ratio, indicating that higher temperatures favour the activation of the decarbonylation pathway.

In Figure 10D, the overall selectivity (S_oxy) of oxygenated intermediates (including alcohol, acid, ester and aldehyde) is depicted. Except for Ni₂P-Ni₁₂P₅/HY and Ni₃P-Ni₁₂P₅/Al₂O₃, which exhibited superior S_oxy values in a lower temperature range, a higher activity of catalysts was prone to show a low oxygenated intermediates formation. These intermediates were further deoxygenated through HDO or decarbonylation pathways during the reaction. These findings are attributed to factors such as Ni site electron properties, the Ni site’s surface density, Ni₂P crystalline size, and synergetic effects between the Ni site, oxygen vacancy, or acid site.

As one of the important factors crucial for the activity of metal phosphide (MP) for hydrodeoxygenation reactions, the metal–phosphorus ratio is investigated. Primarily, phosphorus sites exert the metal sites’ ligand effect by modifying the metal cation’s electron density. This modification facilitates hydrogen segregation, enhancing the catalytic activity of the metal phosphides. Furthermore, metal phosphides may contain phosphorus–oxygen–hydrogen (P-OH) groups, which act as moderately acidic sites. These sites are well-suited to promote hydrogenation reactions, further contributing to the overall catalytic activity of the metal phosphides [118,130].

The impact of the ratio of M/P (metal to phosphorus) was initially explored by Oyama and his co-workers using NiP catalysts with silica support for hydrodenitrogenation (HDN) and hydrodesulphurisation (HDS) reactions [118]. Their study revealed that when the reduction reaction happens, catalysts with excess phosphorus experience a loss of phosphorus content. The HDS reactions were found to be less affected by the Ni/P ratio but played a more significant role in HDN reactions. The phosphorus content profoundly triggered both the activity and stability of HDS, reaching optimum levels starting from 1/2 Ni/P ratio (eventually reaching a 2/1 Ni/P ratio).

At less phosphorus loading, a combination of Ni₁₂P₅ and Ni metal phases was obtained, while at higher phosphorus concentrations, the active phase of Ni₂P was obstructed by a surplus of phosphorus. After selecting an initial M/P ratio, the eventual M/P ratio is influenced by factors such as the reduction temperature (if temperature-programmed reduction is used) and the type of support employed [109]. The composition of the NiP phase depends on both the molar ratio of phosphorus to metal P/M used during synthesis and the type of support material employed.

In silica-supported catalysts, Ni₂P has demonstrated higher activity compared to the HDO of various reactants and phases of Ni₁₂P₅, such as dibenzofuran [79], methyl oleate [131], and methyl laurate [80]. Achieving this stoichiometry requires the use of excess phosphorus and a Ni/P molar ratio lower than the stoichiometric ratio [80,132]. Lowering the Ni/P ratio enhances the reactivity in the HDO of methyl palmitate, leading to increased formation of POx groups. This is explained by the accelerated conversion of methyl palmitate facilitated by acid-catalysed hydrolysis [133]. In the case of palmitic acid HDO using a catalyst based on activated carbon-supported nickel phosphide, a synergetic effect between Ni₁₂P₅ and Ni₂P has been observed [134]. Overall, nickel phosphide has emerged as one of the HDO superior active phases with different reactants, leading to extensive studies on the influence of the M/P ratio for NixPy-based catalysts.

4. Conclusions and Outlook

Vegetable oils must undergo a variety of other processes in addition to hydrotreatment, such as hydrogenation, decarboxylation, and carbonylation, in order to carry out these reactions efficiently. High temperature and pressure are also required for heterogeneous noble or transition metal catalysts to remove oxygen from triglycerides and free fatty acids with great selectivity and to produce the desired hydrocarbons in the C₁₅–C₁₈ range. Among them, sulphided Co-Mo, Ni-Mo, and Ni-W demonstrated a fair degree of efficacy in the deoxygenation and hydrogenation of triglycerides. The compounds have previously been employed in the hydrodesulphurisation process to remove sulphur from petroleum compounds. However, there were significant health dangers because a sulphiding agent that was highly toxic had to be added in order to sustain catalytic activity. For this reason, there has been an increased interest in the generation of biofuels using non-sulphided catalysts, such as Pd, Ni, Ru, Ir, Os, and Rh supported on alumina, silica, and activated carbon. Additionally, certain bimetallic catalysts and alloys have been studied for the hydrotreatment of other vegetable oils.

From nickel-based catalysts, NiP exhibited simultaneous metallic and acidic properties, favouring deoxygenation reactions with low levels of cracking and methanation with the help of its ligand effect and ensemble on metal sites. The comparison of NiP and MoP revealed a low conversion rate of NiP but consumed less hydrogen and had a tendency for specific decarboxylation/decarbonylation pathways than MoP. The addition of suitable supports such as γ-Al₂O₃ to NiP proved its advantages in terms of a large specific surface area, intrinsic acidity, and good thermal stability. In contrast, NiC as a hydroprocessing catalyst for triglycerides and fatty acids had a limiting characteristic, but when supported over Al₂O₃, it exhibited better catalytic activity for the deoxygenation pathway as well. Also, XRD patterns showed a drastic decline in mean pore diameter, surface area, and total pore volume of NiC, NiS, and NiP catalysts after catalyst activation, which indicates a shift in the active phase along with NiO activation. In the case of spent coffee oil hydrotreatment, there was a tendency in catalyst activity in the order NiC > NiP > NiS, in which NiC promoted cracking reaction but with high yields of gasoline products and NiS with lesser conversion rate, while eliminating methanation and cracking reactions. Regarding selectivity to the decarbonylation route of the deoxygenation process, it declined in the order NiS > NiC > NiP, whereas NiS-catalysed products contained a decent amount of aromatics with NiP and NiC only at a low level.

Among the several transition metal catalysts, palladium catalyst-supported over activated carbon has illustrated the highest activity towards the hydrodeoxygenation pathway. From the comparison of Pd/C and NiMo/γ-Al₂O₃ catalysts’ performance on hydroprocessing of palm fatty acid distillate, the Pd/C showed higher diesel yield than NiMo/γ-Al₂O₃. However, for degummed crude palm oil hydrotreatment NiMo/γ-Al₂O₃ exhibited better catalytic activity compared to Pd/C catalyst. Another study was conducted on the production of hydrocarbon fuel from sunflower oil with the combination of a homemade Pd/C and Pt/Ni-Al₂O₃ to enhance the selectivity of the decarboxylation route of the hydrodeoxygenation process for less hydrogen consumption. And this catalytic system reached a 71% conversion for decarboxylation and overall, 86% of vegetable oil conversion into hydrocarbons. This specific method allowed the elimination of the need for an external hydrogen supply and avoided any side reactions such as water gas shift, methanation, and light hydrocarbon formation. However, with spent coffee grounds hydrotreatment, Pd/C enhanced the hydrocracking pathway with higher gasoline yield, but NiMo/γ-Al₂O₃ showed higher isomerisation activity with less olefin content in the liquid product.

The application of commercial CoMo/Al₂O₃ catalyst in hydrotreatment a blend of cottonseed oil and desulphurised diesel proved its long-term activity achieving 70% conversion during the first 100 h and then reaching a steady state after this period. The employment of the catalyst was in a diameter of 1.4 mm in cylindrical form in order to enhance the production of ultra-deep desulphurised diesel. Regarding monitoring the catalyst activity, it steadily decreases until 300 h of reaction time experiencing a loss of almost 13% conversion from its initial value.

The compounds of transition metal phosphides (TMP) have gained interest as highly active hydroprocessing catalysts, as compositions of phosphorous and transition metals which show both physical hardness like in ceramics and electronic properties like in metals. Notably, they had an advantage over other metal catalysts by cost-effectiveness and environmental friendliness, and they also exhibited high activity for hydrogenation reactions. From the families of TMP catalysts, among the monometallic compositions from phosphorus-rich MP₁₅ to metal-rich M₄P, metal-rich phosphide, and bimetallic (M)_x(M’)_yP_z metal phosphides possessed good thermal and electrical conductivity and demonstrated decent thermal and chemical stability. The crystalline structure of phosphides is illustrated in triangular prism forms of metal atoms encompassed by phosphorus atoms. The most interesting part about these catalysts is the fact that they commonly can perform bifunctional properties, including both hydrogenation and acid functionalities. They are able to display these bifunctional characteristics of acidity and hydrogenation capacities thanks to positively charged transition metals acting as Lewis acid sites and phosphate–oxygen–hydrogen chains behaving as Bronsted sites. Among various phosphides supported on silica, there was activity in the order Ni₂P > Co₂P > Fe₂P > WP > MoP, in which Ni₂P/SiO₂ exhibited the highest conversion rate thanks to low acidity or reducibility of SiO₂ support. Lastly, the impact of the phosphorus ratio on the activity of TMP catalysts is crucial for determining the right modification for hydrogen dissociation. However, in actual experiments, it depended on feedstock type and support. For example, in dibenzofuran, methyl laurate, and methyl oleate hydrotreatment, Ni₂P has shown higher activity than Ni₁₂P₅ but with palmitic acid HDO supported on activated carbon, both catalysts went through a synergetic effect with one another.

Overall, the development of efficient heterogeneous catalysts for the hydrotreatment of vegetable oils remains open-ended demanding the need for specific approaches to invent new methodologies and noble properties of advanced catalysis in the future.