Upgrading of Light Bio-oil from Solvothermolysis Liquefaction of an Oil Palm Empty Fruit Bunch in Glycerol by Catalytic Hydrodeoxygenation Using NiMo/Al₂O₃ or CoMo/Al₂O₃ Catalysts

Abstract

Hydrodeoxygenation (HDO) of bio-oil derived from liquefaction of a palm empty fruit bunch (EFB) in glycerol was investigated. To enhance the heating value and reduce the oxygen content of upgraded bio-oil, hydrodeoxygenation of light bio-oil over Ni- and Co-based catalysts on an Al₂O₃ support was performed in a rotating-bed reactor. Two consecutive steps were conducted to produce bio-oil from EFB including (1) microwave-assisted wet torrefaction of EFB and (2) solvothermolysis liquefaction of treated EFB in a Na₂CO₃/glycerol system. The HDO of as-prepared bio-oil was subsequently performed in a unique design reactor possessing a rotating catalyst bed for efficient interaction of a catalyst with bio-oil and facile separation of the catalyst from upgraded bio-oil after the reaction. The reaction was carried out in the presence of each mono- or bimetallic catalyst, namely, Co/Al₂O₃, Ni/Al₂O₃, NiMo/Al₂O₃, and CoMo/Al₂O₃, packed in the rotating-mesh host with a rotation speed of 250 rpm and kept at 300 and 350 °C, 2 MPa hydrogen for 1 h. From the results, the qualities of upgraded bio-oil were substantially improved for all catalysts tested in terms of oxygen reduction and increased high heating value (HHV). Particularly, the NiMo/Al₂O₃ catalyst exhibited the most promising catalyst, providing favorable bio-oil yield and HHV. Remarkably greater energy ratios and carbon recovery together with high H/O, C/O, and H/C ratios were additionally achieved from the NiMo/Al₂O₃ catalyst compared with other catalysts. Cyclopentanone and cyclopentene were the main olefins found in hydrodeoxygenated bio-oil derived from liquefied EFB. It was observed that cyclopentene was first generated and subsequently converted to cyclopentanone under the hydrogenation reaction. These compounds can be further used as a building block in the synthesis of jet-fuel range cycloalkanes.

1. Introduction

Conversion of lignocellulosic biomass into chemicals, fuels, and building blocks for material synthesis such as dehydration of sugars (e.g., glucose, fructose, and xylose) into furans,¹ hydrogenation of furans into fine chemicals, 2,5-furandicarboxylic acid (FDCA),² and cellulose-based materials³ has been intensively studied during past decades regarding the sustainable development goals. Apart from that, liquefaction of biomass has been recently taken into consideration as it is an environmentally friendly route to convert biomass into crude bio-oil and carbon material. The bio-oil as a liquid product can potentially replace petroleum fuels as a renewable energy source, while carbon material can also be upvalued to a precise functional material for various applications such as carbon fibers, carbon clothes, supercapacitors, carbon-based batteries, carbon quantum dots, etc. In thermal liquefaction conversion of biomass, solvothermolysis in the presence of a solvent has been reported to have advantages due to relatively lower operating temperatures (100–200 °C), reduction of tar formation via a cross-linked reaction,⁴ and enhancement of hydrogen donation to the hydrocarbon (HC) structure of the bio-oil product.⁵ Various solvents have been studied for their influence on the dissolution of the biomass matrix during liquefaction including ethylene glycol, poly(ethylene glycol), ethanol, acetone, n-dodecane, and phenol.⁶⁻⁸ Among all solvents investigated, glycerol is an attractive, renewable, and sustainable biosolvent derived as a byproduct from biodiesel production.⁹ Apart from increasing the dissolution effectiveness of biomass constituents, solvents additionally play a vital role as a hydrogen donor when the hydrogenation reaction is performed at a moderate pressure by improving the hydrogen transfer toward the biomass structure and thus boosting up the liquid oil yield with oxygen reduction. The use of glycerol as a hydrogen-donor solvent is a striking means as it is a biobased and renewable solvent relative to polyhydric alcohols. According to the previous studies of bio-oil derived from biomass via thermochemical conversion, the quality of bio-oil was not good enough to be used directly after the conversion process.¹⁰ Therefore, it is necessary to improve the quality of bio-oils similar to that of petroleum fuels by reducing oxygen content, enhancing heating value, and improving physical properties.

Analogous to the hydrotreatment of petroleum fuels, upgrading of biomass liquefied bio-oil using hydrodeoxygenation (HDO) has been intensively studied. HDO is a series of reactions that occur during thermohydrolysis biomass conversion. The HDO reactions mainly include dehydrogenation, decarboxylation, decarbonylation, dehydration, isomerization, and hydrogenolysis of C–O and C–O–C bonds.¹¹ Hydrogenation of carbonyls and C=C bonds or C–C bond cleavages during decarbonylation/decarboxylation is the reaction occurring at metal sites. Hence, the type of metallic sites (e.g., monometallic and bimetallic sites) greatly influences the HDO product selectivity, while the solid acid support facilitates the dehydration reaction. Under severe HDO conditions, the stability of metal particles is additionally critical for metal site selection of the designed catalyst since the sintering and leaching of metals often occur due to the high organic and water content in these reactions. This obstacle is overcome by the use of noble-metal catalysts. Lately, Pt, Pd, and Ru have been reported to exhibit good catalytic performance in such a process¹²⁻¹⁴ since they are more stable than base metal catalysts. However, the applicability of these precious noble metals is hindered by their high costs.

Recently, transition metals (e.g., Ni, Co, and Mo) have been found to be extremely active catalysts for bio-oil reforming to aromatic hydrocarbons.¹⁵⁻¹⁸ Earlier studies have reported on the catalytic activity of CoMo and NiMo catalysts over different supports (e.g., Al₂O₃, SiO₂, activated carbon, zeolite, etc.) for hydrotreating, hydrocracking, and HDO of vegetable oils, waste cooking oil, and triglyceride model compounds for removing oxygen from these oxygenated feeds.¹⁹⁻²² Very recent studies have found that Ni–W over a SiO₂–Al₂O₃ support substantially enhanced the hydrogenation ability of plant-oil triglycerides to aviation biokerosene compared to molybdenum-based catalyst (CoMo, NiMo) systems.²³ The study demonstrated that sulfided Ni–W/SiO₂–Al₂O₃ favored the enhanced hydrocracking of waste soya oil to kerosene, while Ni–Mo/Al₂O₃ facilitated highly active hydrotreatment of it to produce diesel range hydrocarbons.²⁴ Later, the HDO of guaiacol as a model compound containing mainly bio-oil¹⁷ was studied in a batch reactor at 523 K, 5.5 MPa of H₂ in the presence of NiMo and CoMo biometallic catalysts on the potassium-modified γ-Al₂O₃ supporter. It was reported that potassium could increase the selectivity and enhance the yields of products. The main reasons were that (1) the potassium-modified support presumably shifted selectivity from the demethylation and methyl substitution reactions toward the C_Aromatic–OH bond, (2) it could reduce the acidity of the supporter, and thus (3) it prevented catalyst coking on the catalyst surface.

By far, most of the studies on HDO focus on the reaction of model compounds or synthetic diesel from vegetable oils whose chemical structure of the substrate is known. The aim of this investigation was to study the appropriate factors for upgrading bio-oils from solvothermolysis liquefaction of more complex structural biomass. Therefore, the influence of HDO conditions and catalysts, mono- and bimetallic, on a metal support was investigated. Moreover, the energy ratio that includes the yield and high heating value (HHV) due to the deoxygenation process was used to assess the quality of upgraded bio-oil based on raw biomass for the efficient evaluation of biofuels property.

2. Results and Discussion

2.1. Characterization of Catalysts

The suitable temperature for hydrogen reduction of each catalyst was determined by H₂-temperature-programmed reduction (H₂-TPR) analysis. Figure 1 shows the H₂-TPR profiles of Ni/Al₂O₃, Co/Al₂O₃, NiMo/Al₂O₃, and CoMo/Al₂O₃ catalysts. In the case of Co/Al₂O₃, two distinct peaks at around 400 and 620 °C were observed. The larger peak at 400 °C indicated a higher consumption of hydrogen during reduction, which referred to the principal transformation of CoO into a Co⁰ form. This result was in good agreement with a previous work demonstrating that the main reduction peaks from H₂-TPR of cobalt on ZSM-5 zeolite at temperatures lower than 400 °C were ascribed to the reduction of Co₃O₄ that proceeds via a multistep mechanism (reduction of Co₃O₄ to CoO and then to metallic Co).²⁵ The second peak at ∼620 °C was related to the oxides of cobalt toward an alumina support that were more difficult to reduce due to the stronger interaction between the metal sites and the support.²⁶ For the reduction shoulder peak near 320 °C of Co/Al₂O₃, it was reported that Co₃O₄ is relatively easy to reduce (below 450 °C), while CoAl₂O₄ is very difficult to reduce (above 700 °C). Wang and Chen studied the reduction of pure Co₃O₄ and found that for small loading of Co, only one broad peak exists, which belongs to the stepwise reduction of cobalt oxide via Co³⁺ → Co²⁺ → Co⁰.²⁷ Therefore, the first peak may correspond to the reduction of highly dispersed Co₃O₄ and the second may correspond to a CoAl₂O₄-like phase.²⁸

Figure 1.

H₂-TPR analysis of synthesized catalysts, namely, Co, CoMo, Ni, and NiMo supported on Al₂O₃.

The reduction profile of Ni/Al₂O₃ also showed two peaks at 500 and 650 °C. The major peak at 500 °C was assigned to the transition of NiO to metallic Ni⁰. The strong peak at 650 °C possibly involved the stronger metal–support interaction or the reduction of oxide Ni aluminate forms. In addition, the TPR profile of Ni/Al₂O₃ reports that a large reduction peak at ∼760 °C and a shoulder at around 500 °C were associated with the reduction of NiO intensely interacting and weakly interacting with the alumina support, respectively.²⁹ A study by Maia and colleagues also demonstrated the similar reduction patterns of the Ni/zeolite catalyst and proposed three reduction zones: (a) 430–470 °C, (b) 520–560 °C, and (c) 630–720 °C. The first reduction zone can be attributed to the bulk NiO, and the latter two could be due to the smaller NiO particles. Besides, it was also concluded that Ni²⁺ requires a higher temperature to reduce if it is exchanged with H⁺ on the metal oxide of the support structure.³⁰ Therefore, in this case, it can be suggested that bulk NiO and smaller NiO particles with strong oxide interaction of the support were present in the catalyst. This could be further confirmed by X-ray diffraction (XRD) characterization.

For bimetallic catalyst reduction, the reduction profile of CoMo/Al₂O₃ showed three peaks at 380, 580, and 810 °C. The major peak is assigned to the transition of CoO to Co⁰. It has been observed that the second peak at ∼580 °C in the intermediate-temperature zone is assigned to the hydrogen reduction of MoO₃ (Mo⁶⁺) to MoO₂ (Mo⁴⁺). The reduction temperature ranging from 700 to 850 °C is associated with the complete reduction of Mo(IV) to metallic Mo⁰. A similar report also detected the analogous hydrogen consumption peaks of CoMo/Al₂O₃ in a broad temperature range from ∼350 to 950 °C which are composed of three major reduction peaks at 436, 577, and 812 °C.³¹ A high-temperature region with one dominant peak additionally corresponds to the reduction of oxide Co aluminate forms and deeper reduction of Mo⁴⁺ into metallic Mo⁰.³² The reduction profile of NiMo/Al₂O₃ showed three peaks at 320, 520, and 780 °C. The major peak at 550 °C is assigned to the transition of NiO to Ni⁰. The moderate peak at around 800 °C is assigned to the transition of both MoO₂ (Mo⁴⁺) to MoO (Mo²⁺) and further reduction of Mo²⁺ into a metallic Mo⁰ form. The small shoulder at 320 °C was reported on the reduction of unbounded free NiO in the temperature range of 240 and 260 °C.³³ The introduction of Mo into nickel catalysts modified the reduction process and caused the decay of the peak connected with unbound NiO. Compared with the Ni/Al₂O₃ peak, the reduced nickel peak significantly decreased with the addition of Mo, probably due to their strong interaction.³⁴

The XRD technique was used to explain the crystal morphology of the presence of Co, Ni, and Mo on the Al₂O₃ support. XRD runs recorded at a 2θ angular range of 10–80° are shown in Figure 2. The peaks for Co, Ni, and Mo were observed for the oxide phases prior to the hydrogen reduction. In all synthesized catalysts, XRD peaks of Al₂O₃ were detected at 2θ = 38, 46, and 68° corresponding to a report on the characterization of pure γ-Al₂O₃ that showed the diffraction peaks at 2θ = 37, 40, 46, and 67°.³⁵ As demonstrated in Figure 2A,C, NiO shows the peaks at 2θ = 37, 43, and 63°.³⁶ After reduction of Ni/Al₂O₃ and NiMo/Al₂O₃ under a H₂ atmosphere, NiO peaks were more intense relative to those of calcined catalysts. It was previously reported that the peak intensity increased with an increase in calcination temperature. This was indicated by the enhanced degree of crystallization of NiO.³⁶ Therefore, the calcination temperature is a crucial factor and it should be high enough to control the characteristics of NiO. In Figure 2B, the Co-modified catalyst exhibited peak positions i.e., cobalt(II, III) oxide (Co₃O₄), at 2θ = 32, 37, 45, 58, and 66°.^37,38 Therefore, the XRD pattern in this study was confirmed for the cobalt oxide (Co₃O₄) peak of both Co/Al₂O₃ and CoMo/Al₂O₃ catalysts at 2θ = 32, 37, 45, and 66°. Only the Co/Al₂O₃ catalyst showed noticeable peaks of Co²⁺ and Co³⁺ in Co₃O₄ compound at 2θ = 45 and 66° because of the higher loading of Co (5 wt %) on the support compared with CoMo (2.5 wt %). The XRD patterns matched characteristic peaks of cobalt(II) oxide in the form of Co₃O₄ rather than CoO for both calcined and reduced Co/Al₂O₃ and CoMo/Al₂O₃. Thus, overlapping peaks of Co₃O₄, NiO, and Al₂O₃ were found at 37°. In the case of bimetallic NiMo/Al₂O₃ and CoMo/Al₂O₃ catalysts (Figure 2C,D), XRD peaks of molybdenum oxides (Mo⁶⁺ or MoO₃ phase) were detected at a 2θ of 12.7, 23.5, 25.8, 27.3, and 33.9°³⁹ from calcined catalysts. After hydrogen reduction of the NiMO/Al₂O₃ catalyst, three peaks were prominently observed at 26.1, 37.0, and 53.5°, which correspond to MoO₂ or Mo⁴⁺.⁴⁰This demonstrated the transition of Mo⁶⁺ to Mo⁴⁺ after hydrogen reduction of the calcined catalyst as shown in Figure 2C. For reduced CoMo/Al₂O₃, the weak peaks at 2θ = 26.51 and 27.5° corresponding to the (002) and (1̅12) planes of CoMoO₄, respectively, were observed.⁴¹ On comparing XRD patterns of NiMo and CoMo on the Al₂O₃ support, the higher intensity of Mo peaks was found in NiMo/Al₂O₃, presumably indicating the higher performance of Mo in catalytic reactivity for bio-oil hydrodeoxygenation.

Figure 2.

XRD patterns of as-synthesized and hydrogen-reduced (A) Ni, (B) Co, (C) NiMo, and (D) CoMo supported on Al₂O₃ when ● represents Al₂O₃, ◆ represents Ni oxides, ▲ represents Co oxides, and ■ represents Mo oxides.

Binding energy positions of Ni⁰ and Ni²⁺ on the Al₂O₃ support are at 852.4 and 855.5 eV, respectively, for Ni 2p_3/2, which are consistent with the TPR results. There are NiO, Ni(OH)₂, and NiOOH on the surface of the catalysts after calcination in air and hydrogen reduction shown in Figure 3A,B, respectively. Nevertheless, the X-ray photoelectron spectroscopy (XPS) peak areas of metallic Ni⁰ and Ni²⁺ in the H₂-reduced catalyst were enhanced almost 2.5 times as much as those in the calcined Ni/Al₂O₃ catalyst. The presence of NiO may be due to the oxidation of nickel nanoparticles in the air, while an increased Ni(OH)₂ and NiOOH composition is formed by the reaction of Ni²⁺ with hydrogen gas in the reduction stage. Obviously, it can be seen from the Ni 2p XPS that Ni nanoparticles have been successfully deposited on Al₂O₃. It has been demonstrated that the binding energies of free Ni, NiO intimately contacted with Al₂O₃, and NiAl₂O₄ formation into support matrix are different. Free NiO (Ni 2p_3/2) was observed at 853.6 eV, NiO closely interacted with Al₂O₃ was detected at 856.5 eV, and spinel NiAl₂O₄ was noticed at 857.0 eV.⁴² Nevertheless, the XRD results confirm the existence of NiO instead of the spinel phase of the synthesized Ni/Al₂O₃ catalyst. Although the result from H₂-TPR showed the major peak of transition from NiO to metallic Ni⁰ at 550 °C, the ex situ hydrogen reduction of the Ni/Al₂O₃ catalyst at 500 °C for 1 h in the catalyst preparation was achieved. Therefore, the Ni species in the sample calcined and reduced in a hydrogen atmosphere at 500 °C should be mainly NiO weakly contacted with the Al₂O₃ matrix.

Figure 3.

XPS spectra of (A) calcined and (B) reduced Ni/Al₂O₃ for Ni 2p; (C) calcined and (D) reduced Co/Al₂O₃ for Co 2p; (E) calcined and (F) reduced NiMo/Al₂O₃ for Ni 2p; (G) calcined and (H) reduced NiMo/Al₂O₃ for Mo 3d; (I) calcined and (J) reduced CoMo/Al₂O₃ for Co 2p; and (K) calcined and (L) reduced CoMo/Al₂O₃ for Mo 3d.

In the case of calcined and reduced Co/Al₂O₃ catalyst demonstrated in Figure 3C,D, the Co 2p spectra show the characteristic doublet (Co 2p_3/2 at 782.1 eV) corresponding to Co²⁺ of the Co/Al₂O₃ sample. According to the literature, cobalt oxides (Co²⁺ and Co³⁺) found in Co/Al₂O₃ from the impregnation method after calcination in air corresponded to Co₃O₄ (780.0 eV).⁴³ After hydrogen reduction, a greater intensity of all Co XPS peaks was achieved, approximately 2 times compared with those of calcined Co/Al₂O₃ in air. A distinct peak of Co³⁺ appeared at the low binding energy side of the Co 2p_3/2 component at ∼780 eV in agreement with the Co²⁺ peak at 784 eV (Figure 3D). The XPS analytical results were in good accordance with XRD results as Co₃O₄, which contained Co²⁺ and Co³⁺ in the structure, was observed. Its Co 2p_1/2 counterpart can also be detected in the spectra. The considerable increase in the XPS intensity of reduced Co/Al₂O₃ relative to calcined Co/Al₂O₃ can be attributed to a different, probably more reduced cobalt state.

For the NiMo/Al₂O₃ catalyst, Figure 3E,F demonstrates two peaks at 856.3 and 873.5 eV in the Ni 2p XPS spectra assigned to the spin–orbit split lines of Ni 2p_3/2 and Ni 2p_1/2, respectively. The shakeup satellite structures of Ni 2p_3/2 and Ni 2p_3/2 were indicated by two broad peaks at around 862.0 and 880.0 eV^44,45 as shown in Figure 3G,H. Therefore, NiO, Ni₂O₃, and spinel NiAl₂O₄ forms were present in the NiMo/Al₂O₃ sample due to the NiO–Al₂O₃ interface or NiO–MoO₃ interaction that generates spinel NiMoO₄. In contrast to XRD results, no obvious XPS peak attributed to NiO but Ni₂O₃ and spinel NiAl₂O₄ on the surface of NiMo/Al₂O₃ was observed in a range of 853.5–854.5 eV for Ni 2p_3/2.

In the calcined CoMo/Al₂O₃ sample, XPS spectra of Co 2p are similar to those of monometal Co on Al₂O₃ (Figure 3I,J). For Mo 3d spectra, two prominent peaks of Mo⁶⁺ 3d_3/2 and Mo⁶⁺ 3d_5/2 at binding energies of 235.5 and 232.3 eV, respectively, were observed, representing the presence of Mo oxides⁴⁶ (Figure 3K). Mo⁴⁺ 3d_3/2 appearing at 234.0 eV was also detected as a small peak. After hydrogen reduction, as shown in Figure 3L, two new peaks were observed at 229.4 and 232.6 eV attributed to Mo^δ+ (1 < δ < 4) 3d_5/2 and Mo^δ+ 3d_3/2, respectively.⁴⁷ It was reported that Mo^δ+ species were formed when Mo loading was increased. Relative to the calcined CoMo/Al₂O₃ catalyst, Mo^δ+ 3d_5/2 and Mo⁴⁺ 3d_3/2 peak intensities were more intense after hydrogen reduction, while Mo⁶⁺ substantially decreased. It was moreover observed that the binding energies of Mo⁶⁺ and Mo⁴⁺ states shifted toward lower levels after hydrogen reduction for both NiMo/Al₂O₃ (Figure 3H) and CoMo/Al₂O₃ (Figure 3L). This evidence was additionally confirmed by the XRD patterns of MoO₂ occurrence instead of MoO₃ after hydrogen reduction of NiMo and CoMo catalysts in Figure 2C,D, respectively. This shifting of the binding energy of Mo⁶⁺ and Mo⁴⁺ after hydrogen reduction was possibly due to an increased density of electrons that intensified the electromagnetic shield of the inner shell electrons which led to a decrease in binding energy of electrons.⁴⁰

Scanning electron microscopy (SEM) images of all synthesized catalysts, i.e., Ni/Al₂O₃, Co/Al₂O₃, NiMo/Al₂O₃, and CoMo/Al₂O₃ are shown in Figure 4. The metal and the support were aggregated as crystalline solids and possibly provided a large surface area. These SEM images were comparable with the SEM image of commercial activated alumina from a previous work.⁴⁸ It was observed that the morphology of commercial alumina used as a support in this study showed an unsmooth appearance composed of dense aggregates of lumpy anhedral grains of macrocrystallites. Similarly, the characterization of NiMo/Al₂O₃ and CoMo/Al₂O₃ catalysts by SEM images has shown that the small microcrystallites of metal particles on the alumina surface were estimated to be in a nanosized range.⁴⁹

Figure 4.

SEM images of (A) Co/Al₂O₃, (B) CoMo/Al₂O₃, (C) Ni/Al₂O₃, and (D) NiMo/Al₂O₃ catalysts (1000× magnification) after calcination at 500 °C for 5 h.

Further characterization by transmission electron microscopy (TEM) images revealed that metallic Ni, Co, NiMo, and CoMo nanoparticles were highly dispersed on the surface of the Al₂O₃ support (Figure 5). In contrast to SEM images that illustrate the surface characteristics of the Al₂O₃ support using the scanning of an electron beam at a relatively low magnification (1000×), TEM image analysis was applied to characterize the morphology and active phase of Ni, NiMo, Co, and CoMo on the support at a substantially greater magnification (80 000–300 000×). The particle and size distribution of the active phase of the metal in synthesized catalysts was estimated using the electron transmission technique. TEM images of catalysts showed dark spots of Co, Ni, and Mo nanoparticles. The average sizes of each monometallic particle were observed to be 5.1, 48.3, 10.2, and 12.6 nm for Co/Al₂O₃, CoMo/Al₂O₃, Ni/Al₂O₃, and NiMo/Al₂O₃ catalysts, respectively. Zhang et al.⁵⁰ also observed the average sizes of metal particles to be about 7.0 and 5.3 and 7.0 nm for Ni/HZSM-5 and Co/HZSM-5, respectively, by TEM analyses. Although the resolution of the images affected the characterization of detailed Al₂O₃ shape, it appeared as nanosized regular-shaped grains from TEM images. The crystallite sizes of the as-synthesized and commercial Al₂O₃ were approximated to be 10.0 and 12.5 nm, respectively.^48,51 These values were closely comparable with the crystallite size reported for alumina nanoparticles.

Figure 5.

TEM images of (A) Co/Al₂O₃ (300 000× magnification), (B) CoMo/Al₂O₃ (80 000× magnification), (C) Ni/Al₂O₃ (250 000× magnification), and (D) NiMo/Al₂O₃ (150 000× magnification) catalysts after calcination at 500 °C for 5 h.

The analysis of the Brunauer–Emmett–Teller (BET) surface area of reduced catalysts (Table S1) demonstrated that calcined Ni/Al₂O₃ (1.17 × 10² m² g^–1), Co/Al₂O₃ (1.17 × 10² m² g^–1), NiMo/Al₂O₃ (1.03 × 10² m² g^–1), and CoMo/Al₂O₃ (9.82 × 10¹ m² g^–1) had a lesser surface area compared with the Al₂O₃ support (1.27 × 10² m² g^–1). All calcined catalysts had a mesoporous structure in a range of 7.22–7.90 nm pore diameter, indicating that Ni, Co, and Mo particles, which have larger crystal sizes from TEM images (10.2–48.3 nm), were dispersed on the Al₂O₃ surface without entering into the mesoporous structures of the support. Moreover, nitrogen adsorption–desorption isotherms of all calcined catalysts and the Al₂O₃ support revealed type IV isotherm profiles according to IUPAC (Figure S3). This was due to capillary condensation, which is accompanied by hysteresis. The type IV isotherm typically occurs when the pore width exceeds a certain critical width (for pores wider than ∼4 nm), depending on the temperature and system of adsorption. In this case, the pattern of initial monolayer–multilayer adsorption on the mesopore walls is found and subsequently pore condensation is observed.

2.2. Hydrodeoxygenation of Light Bio-oil (LBO) from Empty Fruit Bunch (EFB) Liquefaction

2.2.1. Physical and Chemical Compositions of Upgraded Bio-oil

After improving the quality of bio-oil by HDO in the presence of various mono- and bimetallic catalysts on the Al₂O₃ support at different temperatures, the physical appearance of raw bio-oil (Figure 6A), especially the color, was changed, while the viscosity was observed to be reduced as shown in Figure 6B. It can be seen that the characteristics of bio-oil were enhanced after HDO with the presence of Ni/Al₂O₃, Co/Al₂O₃, NiMo/Al₂O₃, and CoMo/Al₂O₃ catalysts at 350 °C (Figure 6B(2, 4, 6, and 8)) from which the obtained bio-oil was clearer than those at 300 °C (Figure 6B(1, 3, 5, and 7)). Therefore, it could be concluded that upgrading liquefied bio-oils from EFB at 350 °C gave a better bio-oil appearance comparable to petroleum fuels. Among all catalysts and temperatures tested, the HDO of light bio-oil with the NiMo/Al₂O₃ catalyst at 350 °C yielded the clearest upgraded bio-oil as illustrated in Figure 6B6. This result was in good accordance with hydrodeoxygenation and hydrocracking of crude microalgae bio-oil to gasoline, kerosene, and diesel oil by the commercial NiMo/Al₂O₃ catalyst at 250 °C,⁵² and the deoxygenated products were clearly analogous to petroleum-like fuels.⁵³

Figure 6.

Physical appearance of (A) raw light bio-oil (LBO) and heavy bio-oil (HBO) from solvothermolysis liquefaction of EFB and (B) upgraded light bio-oil from the HDO reaction using different reaction times and catalysts: (1) Ni/Al₂O₃ 300 °C, (2) Ni/Al₂O₃ 350 °C, (3) Co/Al₂O₃ 300 °C, (4) Co/Al₂O₃ 350 °C, (5) NiMo/Al₂O₃ 300 °C, (6) NiMo/Al₂O₃ 350 °C, (7) CoMo/Al₂O₃ 300 °C, and (8) CoMo/Al₂O₃ 350 °C with 2 MPa hydrogen for 1 h. (Photograph courtesy of Chutanan Muangsuwan. Copyright 2020).

In terms of chemical compositional analysis of upgraded bio-oil, Fourier transform infrared (FT-IR) spectroscopy was performed for the qualitative analysis as shown in Figure 7. The FT-IR peak assignment of bio-oil is additionally shown in Table 1. The results showed that all deoxygenated bio-oils using different catalysts including Ni/Al₂O₃, Co/Al₂O₃, NiMo/Al₂O₃, and CoMo/Al₂O₃ contained the same functional groups as the blank without deoxygenation. However, there are considerable differences of peak intensities among all catalysts. The peak no. 1 near 3350 cm^–1 represents the O–H stretching, indicating polymeric O–H, alcohols, and phenol found in bio-oil. The peak no. 2 is assigned to CH and CH₂ stretching of alkanes near 1925 and 2960 cm^–1, respectively. Another dominant peak (no. 3) is near 2355 cm^–1, representing the presence of a −C–O group, the peak no. 4 in the range of 1602–1656 cm^–1 represents the −C=C– group of alkenes, and the peak no. 5 at 1043 cm^–1 is attributed to the −C–O group of phenols, ester, and ethers.⁵⁴ Compared to the blank without HDO, the bio-oil from HDO has higher peak intensities for the absorbance at 2925 and 2960 cm^–1 attributed to higher CH and CH₂ contents in bio-oil due to hydrogenation and deoxygenation reactions. Similarly, the greater peak intensity at 1642 cm^–1 of HDO of bio-oil with catalysts shows the reduction of oxygen groups such as OH and C–O, while the amount of −C=C– groups increased substantially. In addition, the presence of FTIR peaks at 1650 and 1750 cm^–1 represents the C=O stretching of ketones, aldehydes, and carboxylic acids. This conclusion is also confirmed by the decline of peaks assigned to hydroxyl groups at around 3350 cm^–1 concerning the deoxygenation of aliphatic alcohols to esters. From the results, liquefied bio-oil from biomass was upgraded by elimination of oxygen atoms from their composition. A similar work has reported the elimination of oxygen atoms from pyrolysis of bio-oil.⁵⁵

Figure 7.

FT-IR analysis of hydrodeoxygenated bio-oil at (A) 300 °C and (B) 350 °C using different catalysts with 2 MPa hydrogen for 1 h.

Table 1. Peak Assignment from FT-IR Spectroscopy of Functional Groups of Bio-oil⁵⁶⁻⁵⁸.

peak	wavenumber (cm–1)	functional group	class of compound
1	3350	O–H stretching	polymeric O–H
2	2960, 2925, 2850	CH, CH2 stretching, C–H stretching	alkanes
3	2354.66	C–O stretching	phenol, alcohols, esters, ether
4	1690–1759, 1700	C=O stretching	ketones, aldehydes, carboxylic acids, mono-alkyl ester
	1602–1656	C=C stretching	alkenes
	1515	C=C ring stretching	aromatics
	1282, 1205	C–O–C stretching	phenol, ester, ethers
	1240, 1260	C–H stretching	aromatics
5	1195, 1110, 1043	C–O stretching	phenol, ester, ethers
	885	C–H deformation	aromatics
	750	adjacent C–H deformation	aromatics
	696	out-of-plane =CH deformation	alkenes
	615	out-of-plane O–H deformation	polymeric O–H

Among all catalysts tested, the remarkable peak intensity near 1043 cm^–1 (peak no. 5) attributed to the −C–O group of phenols, esters, and ethers is declined in the case of the NiMo/Al₂O₃ catalyst compared with others and the blank for both temperatures (300 and 350 °C) in Figure 7A,B. Moreover, an increase of peak intensity was found to be pronounced at the peak no. 4 in the range of 1602–1656 cm^–1 representing the greater amount of the −C=C– group of alkenes in upgraded bio-oil from NiMo/Al₂O₃ compared with the blank. Furthermore, the smaller ratio of intensity at 1210 to 1656 cm^–1 representing C–O stretching of acids and C=C stretching of alkenes, respectively, was detected. This demonstrated that the acid content in upgraded bio-oil from NiMo/Al₂O₃ catalyst was considerably less than that from other catalysts and the blank when compared with the alkene content. Smaller peaks at the bands at 970 and 1250 cm^–1, assigned to the C–O stretching of phenols and alcohols,⁵⁸ were detected in upgraded bio-oils from the NiMo/Al₂O₃ catalyst. Molybdenum plays an important role in Ni metallic to enhance the reduction of C–O in phenols to aromatics as well as alcohols to either alkenes or alkanes in the presence of hydrogen,⁵⁹ but a less effect was observed for Co metallic. Therefore, from FT-IR qualitative analysis, the NiMo/Al₂O₃ catalyst provided significant improvement of bio-oil deoxygenation relative to other catalysts at both temperatures investigated.

The gas chromatography-mass spectrometry (GC-MS) analysis of bio-oil confirmed the FT-IR absorption of chemical constituents in upgraded bio-oil. Figure 8 illustrates the high qualitative composition found in bio-oil. Approximately 10 compounds from both HDO temperatures of 300 and 350 °C, which have high peak area percentage among all peaks, were observed, including ethanol, 1-propanol, propanoic acid, 2-methyl-cyclopentanone, 2-methyl-cyclopentene-1-one, 3-methyl-cyclopentene-1-one, phenol, 3,4-dimethyl-cyclopentene-1-one, 2,3-dimethyl-cyclopentene-1-one, and 2-methyl phenol as demonstrated in Figure 8A,B. The effect of different catalysts on HDO showed that in the presence of a catalyst, the quantities of lower molecular weight phenolic, ester, and ketone were found compared with the blank without adding the catalyst. At 300 °C, NiMo/Al₂O₃ demonstrated more selective products such as 3-methyl-2-cyclopentene-1-one, 2-methyl-2-cyclopentene-1-one, and phenol (Figure 8A). It seems that demethylation and methylation reactions are compatible at this temperature as 2-methyl phenol and phenol quantities as well as the relative amount of 2,3-methyl-2-cyclopentene-1-one, 3,4-methyl-2-cyclopentene-1-one, 3-methyl-2-cyclopentene-1-one, and 2-methyl-2-cyclopentene-1-one were not significantly different. The results from the catalytic reaction support the XRD and XPS characteristics of the catalyst that contained the majority of MoO₃ (Mo⁶⁺) and MoO₂ (Mo⁴⁺) on the Al₂O₃ support, which selectively catalyze the demethylation of either guaiacol or m-cresol to phenol or benzene.⁶⁰ Although the previous study revealed that MoO₃ is more active than MoO₂ on HDO of biomass-derived compounds, oxygen appearance during HDO could oxidize MoO₂ to a more stable MoO₃ phase that provides higher selectivity and conversion for preferentially cleaving phenolic Ph–OMe bonds over weaker aliphatic Ph–O–Me bonds under low hydrogen pressure and temperature in the range of 320–350 °C.⁶¹

Figure 8.

High abundance of chemicals found in upgraded bio-oil from HDO with various catalysts at (A) 300 °C, and (B) 350 °C with 2 MPa hydrogen for 1 h.

However, when the HDO temperature was increased to 350 °C as demonstrated in Figure 8B, the presence of Mo in NiMo/Al₂O₃ showed a significant influence on demethylation of 2,3-methyl-2-cyclopentene-1-one and 3,4-methyl-2-cyclopentene-1-one to generate 3-methyl-2-cyclopentene-1-one and 2-methyl-2-cyclopentene-1-one, respectively. This proved that as temperature increased, the relatively larger molecules were prone to bond-breaking and reforming to generate lower molecular cyclopentanone structures. In contrast, Ni/Al₂O₃ substantially enhanced methylation of 2-methyl-2-cyclopentene-1-one to 2,3-methyl-2-cyclopentene-1-one with high selectivity. At this temperature, more acids and alcohols were generated, especially from Co/Al₂O₃ and CoMo/Al₂O₃, which led to a worse effect on the bio-oil quality due to high oxygen content and low HHV. This occurrence corresponds to the characteristics of CoMo from XRD and XPS results, which contained a greater composition of MoO₂ (Mo⁴⁺) compared with NiMo, which contained more MoO₃ (Mo⁶⁺) phase. MoO₃ (Mo⁶⁺) is more favorable to selectively catalyze guaiacol and m-cresol to phenol and benzene. Since the bond-dissociation energy analysis of relevant phenolic C–O bonds indicates that the bond strengths follow an order of Ph–OH > Ph–OMe > Ph–O–Ph > Ph–O–Me,⁶¹ therefore MoO₂ (Mo⁴⁺) in CoMo as well as CoMoO₄ seemingly promoted unstable compounds, such as acids, furfural, vanillin, and levoglucosan that were able to be converted into esters, ketones, and saturated phenols⁶² relative to NiMo.

According to Table 2, chemical constituents in bio-oil could be categorized into seven groups, namely, alkanes, acids, alcohols, phenols, esters, aldehydes, and ketones. In upgraded bio-oil without the addition of a catalyst, an equal amount of alcohols and ketones was obtained. Relatively high amount of phenolic compounds was observed due to the breakdown of C–C bonds and ether linkages found in the lignin polymer’s side chains, which was the major constituent in bio-oil.⁶³ Additionally, the majority of G-phenol products in bio-oil derived from guaiacyl accounted for more than 85% of the total phenols in both raw and upgraded bio-oil. The major contributing compound was 2-methyl-phenol. In the case of ketone generation, the larger ketone molecules such as cyclopentenone and cyclopentanone were derived from the pyrolysis or liquefaction of hemicellulose or cellulose. The cyclopentanone and cyclopentenone were initially produced from the degradation of sugar, and subsequently, opened bond recombination occurred, while the smaller ketone molecule formation was more complicated.

Table 2. Chemical Composition in Upgraded Bio-oil from EFB Solvothermolysis Liquefaction.

In fact, the amount of different bio-oil compositions depends on each type of catalyst, and the constituent required in large quantities as a worthy bio-fuel from HDO is alkane. This consequence is also indicated by high C/O ratio and high HHV. Nevertheless, the amount of generated acids and alcohols needs to be concerned as it mainly causes corrosion and deterioration of combustion engines. In this work, Figure 9 demonstrates the relative quantity of each group present in bio-oil products from different catalysts and HDO temperatures. As shown in Figure 9A, the results indicated that the aromatic ketones were the main components from HDO at 300 °C, followed by alcohols, acids, and alkanes from most catalysts. Among ketone products (Table S2), cyclopentanone and cyclopentene were the main olefins found in hydrodeoxygenated bio-oil derived from liquefied EFB. From GC-MS analysis, it was observed that cyclopentene was first generated and subsequently converted into cyclopentanone under the hydrogenation reaction. Cyclopentanone is an important lignocellulosic platform compound derived from degradation products of hydrolysis–dehydration of cellulose and hemicellulose.⁶⁴ In recent literatures,⁶⁵⁻⁶⁸ it was demonstrated that cyclopentanone can be produced in high carbon yield by the aqueous-phase selective hydrogenation of furfural from the subsequent hydrolysis–dehydration of hemicellulose.⁶⁹ This compound has a cyclic structure and can be used as a building block in the synthesis of jet fuel range cycloalkanes.

Figure 9.

Chemical compositions of bio-oils from HDO with 2 MPa hydrogen for 1 h in the presence of different catalysts at (A) 300 °C and (B) 350 °C.

Conversion of 3-hydroxycyclopentanone into 2-cyclopentene-1-one and to cyclopentanone was postulated to occur by hydrogenation. The difference in the solubility of the substrate, intermediates, and products, along with the optimum ratio of metals in the synthesized catalyst, plays a crucial role in achieving better yields.⁷⁰ Moreover, it was reported that the solvent and temperature considerably influenced the selectivity of the hydrogenation product of furfural. In an aqueous solution, high selectivity of furfural conversion into cyclopentanone was achieved at 160 °C (70%) > 140 °C (3%) > 120 °C (0%); however, too high temperature (180 °C) also led to 0% selectivity of cyclopentanone due to further conversion of cyclopentanone into cyclopentanol under a moderate hydrogen pressure (2–4 MPa).⁷¹ Co/Al₂O₃ produced substantially high amount of acids, which was not favorable for bio-oil upgrading. The presence of Mo in bimetallic CoMo/Al₂O₃ and NiMo/Al₂O₃ catalysts gave significantly higher aromatic ketone composition in bio-oil, which was also satisfactory due to a higher HHV and C/O ratio.

The upgraded bio-oil at 350 °C (Table S3) contains aromatic ketones as the majority for Co/Al₂O₃, Ni/Al₂O₃, and NiMo/Al₂O₃ catalysts in respective degrees as shown in Figure 9B. The Ni/Al₂O₃ catalyst in either NiO or NiO(OH) (Ni²⁺) form from the XPS and XRD results promoted the generation of a substantial amount of 2-methyl-phenol, which was possibly due to methylation of phenol. Similar findings were reported on the decrease of guaiacol and 1,2-benzenediol after HDO since they were converted into cyclohexanone and aromatic HCs such as biphenyl, 1,2,3,4-tetrahydro-5,6-dimethyl-naphthalene, hexamethyl benzene, and 1,3-dimethyl-1-cyclohexene via CH₃-substitution, dehydration, or methyltransfer reactions.⁷² An enhancement of H₂ availability in the reaction liquid by augmenting H₂ pressure or the addition of a H₂ donor solvent significantly promotes the transformation of guaiacol into fully deoxygenated products via HDO with high efficiency.⁷³

A significantly higher quantity of alkane was also found in the Co/Al₂O₃ catalyst at this temperature with a smaller amount of acids compared with the upgraded bio-oil at 300 °C. This was a favorable consequence; however, the upgraded bio-oil yield from the Co/Al₂O₃ catalyst (27 wt %) was the lowest among all catalysts. Moreover, upgraded bio-oil from CoMo/Al₂O₃ and NiMo/Al₂O₃ catalysts generated high amount of unidentified substances, which may contain high HHVs with low oxygen content. Accordingly, the C/O ratio and energy ratio, in which the yield and HHV were taken into account, should be considered for the finest selection of HDO conditions for upgrading bio-oil from EFB liquefaction.

Based on the GC-MS results, the HDO of the main product in bio-oil derived from liquefaction of EFB catalyzed by the NiMo/Al₂O₃ catalyst is demonstrated in Figure 10. The main reactions of guaiacol-like compounds included (1) demethylation reactions of the CH₃O group by the hydrogenolysis of the O–CH₃ bond of the methoxy group, or demethoxylation via the cleavage of the C_aromatic–OCH₃ bond, and (2) in the case of phenolic OH-group the first conversion pathway is the hydrogenolysis of the C_aromatic–OH bond, while the second pathway involves the hydrogenation of the initial aromatic ring followed by the OH-group elimination.⁷⁴ The ability of a metallic Ni-based catalyst for a slightly inferior concentration of protons in the aromatic region has been reported, indicating higher activity toward hydrogenation of the aromatic rings in bio-oil, similar to a Ru/C-based catalyst.⁷⁵ In the present study, the appearance of the NiO or NiO(OH) phase from XRD and XPS results also indicated the ability of oxygen adsorption toward the cleavage of aromatic rings promoted by the presence of Mo in the NiMo/Al₂O₃ catalyst. The incorporation of nickel and molybdenum in the synthesized catalyst with a proper Ni/Mo ratio as well as hydrogen partial pressure and temperature substantially influences the ring-opening reaction in the hydrogenation process, attributable to new acid sites associated with the molybdenum species.⁷⁶

Figure 10.

Proposed schematic mechanism of demethylation and the ring-opening reaction of 2,3- and 3,4-dimethylbenzene catalyzed by the NiMo/Al₂O₃ catalyst.

Previous reports on the recyclability study of NiMo/Al₂O₃ and CoMo/Al₂O₃ catalysts stated that the HDO reactivity of NiMo and CoMo on the Al₂O₃ support could be substantially enhanced after the first recycle compared with the fresh catalyst in the presence of hydrocarbons containing high sulfur content.⁷⁷ High stability up to five times of the spent catalyst for oil upgrading at 400 °C, an initial H₂ pressure of 5 MPa, and 800 rpm for 1 h reaction could be obtained. Recovery of metals from spent hydrotreated catalysts could be efficiently done by mineral phase reconstruction using sodium carbonate and carbon powder followed by stepwise extraction. The leaching ability of Al and Mo was modulated by different ratios of substances added, the solid-to-liquid ratio, and temperature and time.⁷⁸

2.2.2. Ultimate Compositional Analysis of Hydrodeoxygenated Bio-oil

The amount of elemental C, H, N, and O, which are components in bio-oil before and after the HDO process using different catalysts and temperatures, is shown in Table 3. The raw bio-oil derived from EFB solvothermolysis liquefaction contained C, H, N, and O contents of 30.8, 9.1, 6.4, and 42.3 wt %, respectively. A higher nitrogen content in crude bio-oil was obtained when compared with raw EFB since the pressurized nitrogen gas was applied during biomass liquefaction (300 °C, 8 MPa N₂ for 1 h). Another carbon yield of approximately 50 wt % was in heavy oil, which was not included in this study.

Table 3. Elemental Analysis and Carbon Yield from Hydrodeoxygenation of Bio-oil at Different Temperatures and in the Presence of Different Catalysts.

			elemental content of light bio-oil (wt %)					elemental ratio
substrate	upgrading condition (temperature, catalysts)		C	H	N	O	S	H/O (by mol)	H/C (by mol)	C/O (by mol)	bio-oil yield (wt %)	carbon yielda (%)
raw EFB	nab		42.25	5.88	1.03	50.75	0.09	1.85	1.67	1.11	na	100
raw bio-oil from EFB liquefaction	na		30.81	9.1	6.36	42.27	na	3.44	3.54	0.97	61.41	44.78
upgraded bio-oil	300 °C	Ni/Al2O3	28.84	9.56	5.9	48.1	na	3.18	3.98	0.80	46	43.06
		Co/Al2O3	27.37	9.62	5.52	55.74	na	2.76	4.22	0.65	44	41.76
		NiMo/Al2O3	25.44	9.66	5.94	41.05	na	3.77	4.56	0.83	45	41.83
		CoMo/Al2O3	26.2	9.43	6.3	49.77	na	3.03	4.32	0.70	41	42.22
	350 °C	Ni/Al2O3	23.96	10.42	4.77	44.1	na	3.78	5.22	0.72	41	37.49
		Co/Al2O3	29.66	10.27	4.02	49.43	na	3.32	4.16	0.80	27	33.42
		NiMo/Al2O3	37.69	10.77	4.96	38.91	na	4.43	3.43	1.29	40	50.83
		CoMo/Al2O3	38.21	10.3	4.01	35.11	na	4.69	3.23	1.45	35	35.48

^aIn this work, the carbon yield was calculated from only the light bio-oil portion and the heavy oil portion was excluded.

^bna = not applicable.

From the results, oxygen reduction was significantly observed from HDO at 350 °C from NiMo/Al₂O₃ and CoMo/Al₂O₃ catalysts indicated by higher H/O and C/O ratios. It was found that the oxygen content decreased from 42.3 to 38.9%, corresponding to 7.9% oxygen removal after the HDO catalyzed by the NiMo/Al₂O₃ catalyst at 350 °C. The decreased oxygen content is the result of decarbonylation/decarboxylation to generate steam, CO, and CO₂.⁷²⁷² In addition, it was found that after the HDO process in all conditions, the hydrogen content was increased. This shows that the hydrogenation/hydrogenolysis of bio-oil occurs during the HDO process.

The H/O, H/C, and O/C ratios can be used as the indicators of the hydrogenation and deoxygenation processes for upgrading oxygenated oil for the replacement of petroleum fuels. Considering the conventional fuels such as diesel and fatty acid methyl ester,⁷⁹ the ultimate goal of hydrogenation is to achieve the liquid product with O/C and H/C ratios ranging from 1.5 to 2.0.²³ These values correspond to 0.5–0.7 of the C/O ratio. In Table 3, the results in this study showed that all of the upgraded bio-oils from HDO of EFB-derived bio-oil have higher C/O ratios than the target due to successful oxygen reduction and hydrogenation. In particular, HDO catalyzed by CoMo/Al₂O₃ and NiMo/Al₂O₃ catalysts at 350 °C gave the upgraded bio-oil with C/O ratios at 1.45 and 1.29, respectively, which were substantially greater than those at other conditions. These values corresponded to O/C ratios of 0.69–0.78, which were still higher than those of upgraded liquefied bio-oil from cornstalk by catalytic hydrodeoxygenation using bimetallic ammonium nickel molybdate at different temperatures (280–370 °C) under supercritical ethanol with initial 4 MPa H₂, 60 min, giving H/C and O/C molar ratios of upgraded bio-oil in the range of 1.19–1.30 and 0.05–0.15, respectively, depending on temperatures.⁸⁰ Furthermore, upgraded bio-oil from the NiMo/Al₂O₃ catalyst at 350 °C gave the highest carbon yield of 50.83%; therefore, it was considered the most promising catalyst in this work.

2.2.3. HHV of Hydrodeoxygenated Bio-oil

The HHVs of the bio-oil from HDO using four different catalysts at 300 and 350 °C are shown in Table 4. It was found that the HHV values from HDO at 300 °C catalyzed by Ni/Al₂O₃, Co/Al₂O₃, NiMo/Al₂O₃, and CoMo/Al₂O₃ catalysts were 38.77, 39.88, 39.15, and 39.57 MJ kg^–1, respectively. Among all catalysts investigated at 300 °C, deoxygenated bio-oil catalyzed by the Co/Al₂O₃ catalyst gave the highest HHV; however, the value was insignificantly different compared with those of the other catalysts.

Table 4. High Heating Value (HHV) and Energy Ratio of All Upgraded Bio-oils from Hydrodeoxygenation at Different Temperatures and in the Presence of Different Catalysts^a.

				light bio-oil			heavy bio-oil
biomass	upgrading condition (temperature, catalysts)		HHV of raw material (MJ kg–1)	yield (%)	HHV (MJ kg–1)	energy ratio (r)	yield (%)	HHV (MJ kg–1)	energy ratio (r)	total energy ratio (r)
raw bio-oil from EFB liquefaction	nab		13.02	61	36.67	1.73	15.39	40.56	0.48	2.21
upgraded bio-oil	300 °C	Ni/Al2O3	13.02	46	38.77	1.37	15.39	40.56	0.48	1.85
		Co/Al2O3	13.02	44	39.88	1.35	15.39	40.56	0.48	1.83
		NiMo/Al2O3	13.02	45	39.15	1.35	15.39	40.56	0.48	1.83
		CoMo/Al2O3	13.02	41	39.57	1.25	15.39	40.56	0.48	1.73
	350 °C	Ni/Al2O3	13.02	41	40.07	1.26	15.39	40.56	0.48	1.74
		Co/Al2O3	13.02	27	42.53	0.88	15.39	40.56	0.48	1.36
		NiMo/Al2O3	13.02	41	40.87	1.29	15.39	40.56	0.48	1.77
		CoMo/Al2O3	13.02	35	41.33	1.11	15.39	40.56	0.48	1.59

^aNote: The energy ratio was calculated based on the HHV of raw EFB at 13.02 MJ kg^–1 for all cases.

^bna = not applicable.

Bio-oil HDO at 350 °C using Ni/Al₂O₃, Co/Al₂O₃, NiMo/Al₂O₃, and CoMo/Al₂O₃ catalysts gave superior HHVs of 40.07, 42.53, 40.87, and 41.33 MJ kg^–1, respectively, compared with those under HDO at 300 °C. At 350 °C, deoxygenated bio-oil from HDO in the presence of Co/Al₂O₃ had the highest HHV value, but the lowest yield was obtained after the deoxygenation reaction. This corresponded to the previous report stating that the yield of the product decreased when the HDO temperature and time of heating increased.⁸¹ In the presence of a catalyst, more gaseous products are generated during the deoxygenation reaction at high temperature and thus a significantly lower yield was obtained. A higher HHV of bio-oil was achieved as a result of oxygen reduction in the upgraded bio-oil. During HDO, a promising catalyst could lead the favorable reactions that allow C=C bonds (614 kJ mol^–1) and C–O bonds (358 kJ mol^–1) to break apart by hydrotreatment, while new C–C bonds (348 kJ mol^–1) and C–H bonds (413 kJ mol^–1) are formed during hydrogenolysis.⁸² Long-chain alkane products from HDO in the presence of the Co/Al₂O₃ catalyst at 350 °C generated the upgraded bio-oil with a substantially greater HHV as shown in Table 4. Table S4 shows the comparison of O/C and H/C ratios, HHVs, and main products from hydrodeoxygenation (HDO) of bio-oil using Ni- and Co-based catalysts from this study and previous studies.

A higher HHV of the bio-oil implies a higher portion of the energy in the feed that turns into the bio-oil. However, after deoxygenation of light bio-oil, a lower yield was obtained due to the loss of low-molecular-weight compounds into gaseous products. Thus, a more complete measurement of the process in delivering energy from biomass to the bio-oil is the energy ratio (r). This value includes the mass yield and HHV of bio-oil and is calculated relative to the HHV of raw material. According to the results, Table 4 shows that the NiMo/Al₂O₃ catalyst in HDO at 350 °C was the most promising catalyst and condition that produced deoxygenated bio-oil with markedly high energy ratios. Moreover, this condition provided the greatest carbon yield as well as the highest ratios of C/O, H/C, and H/O (Table 3), indicating the good quality of bio-fuel for combustion engines. Previous reports also confirmed that metallic Ni and Co are active in C–C bond hydrogenolysis^17,83 and thus give rather a worse effect since the increased hydrogen consumption and the decreased carbon yield were obtained. Moreover, sole metallic Ni and Co suppressed the hydrogenation of benzene rings. Therefore, the interaction of a second metal promoter as a bimetallic catalyst such as Mo has been proposed.⁸⁴ The HDO with the NiMo/Al₂O₃ catalyst at 350 °C produced the highest quality of upgraded bio-oil derived from EFB liquefaction with the least release of SO_x and NO_x with high HHV.

3. Conclusions

Upgrading bio-oil from solvothermolysis liquefaction of a palm empty fruit bunch by catalytic HDO in the presence of monometallic Ni and Co and bimetallic NiMo and CoMo over an alumina support was investigated at different temperatures of 300 and 350 °C. From the results, HDO catalyzed by NiMo/Al₂O₃ at 350 °C gave the highest bio-oil yield. Apart from that, HDO in the presence of the NiMo/Al₂O₃ catalyst at 350 °C generated bio-oil with the greatest energy ratio, carbon yield, and highest ratios of C/O, H/C, and H/O, indicating the good quality of bio-fuel for combustion engines with the least release of SO_x and NO_x. Cyclopentanone, cyclopentene and related compounds were the main olefins found in hydrodeoxygenated bio-oil derived from liquefied EFB. These compounds are building blocks for the synthesis of jet fuel range cycloalkanes. Interestingly, these catalysts work effectively without sulfidation processes that could incorporate sulfur into bio-oil composition, which leads to environmental problems. The findings provide a competent alternative for thermochemical conversion of wet biomass into liquid fuels in an energy-efficient approach.

4. Experimental Section

4.1. Materials

To produce bio-oil, a palm empty fruit bunch (EFB) was selected as a raw material for solvothermolysis liquefaction. EFB is the most abundant solid residue accounting for 20 wt % of fresh palm bunch including fruits. EFB biomass is produced at approximately 4.42 t ha^–1 year^–1 after oil extraction at palm oil mills.⁸⁵ EFB was obtained from Chumporn Palm Oil Industry Public Company Limited (CPI), Thailand. The precursors for catalyst preparation, namely, Ni(NO₃)₂·6H₂O, Co(NO₃)₂·6H₂O, (NH₄)₆Mo₇·4H₂O, and the alumina support were purchased from Sigma-Aldrich.

4.2. Microwave-Assisted Torrefaction and Solvothermolysis Liquefaction of EFB

The EFB sample was initially washed with tap water and dried in an oven at 105 °C for 12 h before crushing and milling to a consistent particle size of +50/–200 mesh by sieving. The moisture content of the feedstock was determined by a moisture analyzer (Ohaus MB120). Prior to biomass liquefaction, 2 g of biomass was mixed with 20 mL of glycerol and subjected to microwave irradiation (Anton Paar, Austria) at 800 W for 20 min with a heating-up duration of 2 min. Five reaction vessels were used simultaneously. Microwave energy is able to couple with the molecules in a pretreatment mixture to provide more volumetric and energy-efficient internal heating of a substrate.⁸⁶

After microwave pretreatment, the solvothermolysis liquefaction was performed in a 500 mL reactor. In the experiment, 100 mL of the microwave-treated slurry and 1 g of the Na₂CO₃ catalyst (1 wt %) were mixed in the reactor. The reactants were agitated using an external stirrer at 150 rpm equipped with a magnetic seal drive. The temperature was controlled at 300 °C (heating rate was 10 °C min^–1), and nitrogen gas was applied to set the initial pressure at 8 MPa. After 1 h reaction was completed, the reactor was cooled down to room temperature by a cooling water system.

Heavy bio-oil and light bio-oil were generated after biomass liquefaction, and they were separated by a separation funnel. Ultimate analysis was performed using a CHNS/O analyzer, and the high heating value (HHV) of light bio-oil was measured using a bomb calorimeter. The chemical composition of bio-oil was analyzed by gas chromatography-mass spectrometry (GC-MS) after liquid–liquid extraction using toluene as an extractant. The methods of preparation and analysis are reported elsewhere.¹⁰ Light bio-oil was further upgraded using catalytic HDO.

4.3. Catalysts Preparation by the Wet Impregnation Method

Wet impregnation of monometallic and bimetallic catalysts on the alumina support (Al₂O₃) was selected for catalyst synthesis in this study. Metal precursors for monometallic catalysts Ni/Al₂O₃ and Co/Al₂O₃ were Ni(NO₃)₂·6H₂O (290.79 g mol^–1) and Co(NO₃)₂·6H₂O (291.03 g mol^–1), respectively. At 5% w/w loading based on the support, Ni(NO₃)₂·6H₂O or Co(NO₃)₂·6H₂O was weighed based on Al₂O₃ and dissolved in 5 mL of deionized (DI) water. The solution was poured dropwise over the dry Al₂O₃ support in a crucible and dried at 100 °C for 12 h. To prevent the particle agglomeration and crystallinity change, the calcination condition selected in this study was 500 °C for 5 h in air.⁸⁷ After that, H₂-TPR analysis was used to find the optimal temperature for catalyst reduction. Before utilization in the HDO, the calcined catalysts were reduced under a H₂ stream with the flow rate of 80 mL min^–1 at 500 °C for 1 h.

For bimetallic catalysts, NiMo/Al₂O₃ and CoMo/Al₂O₃, an additional Mo precursor was used, which was (NH₄)₆Mo₇·4H₂O (1235.86 g mol^–1). To prepare the Ni–Mo/Al₂O₃ catalyst at the weight ratio of Ni/Mo at 1:1, Ni(NO₃)₂·6H₂O was mixed with (NH₄)₆Mo₇·4H₂O to maintain 5 % w/w based on the Al₂O₃ support and subsequently dissolved in 5 mL of deionized water. In the case of CoMo/Al₂O₃ synthesis, a Co(NO₃)₂·6H₂O precursor was mixed with (NH₄)₆Mo₇·4H₂O at 5% w/w based on the Al₂O₃ support and then dissolved in 5 mL of deionized water. The solution was added dropwise over the dry Al₂O₃ support and followed by calcination and H₂ reduction before use according to the procedure aforementioned.

4.4. Catalytic HDO of Light Bio-oil from Solvothermal Liquefaction of EFB

The upgrading of bio-oil derived from solvothermolysis liquefaction of EFB was performed in a 100 mL rotating-bed reactor. Into the reactor, 20 g of light bio-oil and 2 g of the synthesized catalyst (10 wt % catalyst loading) were added. The reaction system was flushed with H₂ gas several times to ensure that the oxygen was removed. The catalytic HDO of bio-oil was then initiated under 2 MPa of initial H₂ pressure at 300 and 350 °C with a heating rate of 12 °C min^–1. The rotation speed of the catalyst bed was kept constant at 250 rpm, and the reaction time was 1 h. After the reaction, the upgraded bio-oil was cooled down using a refrigeration system.

4.5. Characterization of Synthesized Catalysts

The physical and morphological characterization of synthesized catalysts was performed by scanning electron microscope (SEM) and transmission electron microscopy (TEM, JEM-2100/HR (200 kV), resolution point 0.23 nm and lattice 0.14 nm, single tilt and double tilt (±30°) holder). X-ray diffraction (XRD) was applied for the identification of metal crystallization using an XRD spectrometer equipped with a Cu tube (XRD, Malvern PANalytical, Model Aeris, Netherland) with a 0.02° min^–1 scan rate from 2θ = 10 to 80°. The suitable temperature for catalyst reduction was analyzed by H₂-temperature programmed reduction (H₂-TPR) using 200 mg of the catalyst at 20 mL min^–1 flow rate of Ar at 400 °C for 40 min. Then, the temperature was decreased to 40 °C, and subsequently, the catalyst was reduced in a 5% H₂/Ar stream with a heating rate of 10 °C min^–1 up to 800 °C. The valence electron of elemental atomic surface and the chemical binding and composition on the surface of the synthesized catalyst were analyzed by X-ray photoelectron spectroscopy (XPS) (AXIS Ultra DLD model, Kratos Analytical Ltd., U.K.). Nitrogen adsorption and desorption isotherms (280 °C) as well as the BET surface area, pore size diameter, and pore volume distribution of the support and reduced catalysts were analyzed (BELSORP-mini II model, BEL, Japan) using 150 °C, 6 h degas conditions.

4.6. Characterization of Upgraded Bio-oil from the HDO Process

Functional groups in bio-oil products were analyzed by Fourier transform infrared (FT-IR) spectroscopy using Nicolet 6700 model, Thermo Fisher Scientific. The measurement was performed in the wavenumber ranging from 4000 to 400 cm^–1 for 100 scan numbers with 4 cm^–1 resolutions. Compositional analysis of bio-oil as the mass percentage of C, H, O, and N was performed by a CHNS/O analyzer (2400 series II, PerkinElmer). A bomb calorimeter (PARR 1261, USA) was operated for the measurement of the higher heating value (HHV) of bio-oil when 1 g of benzoic acid was used as a heat of combustion standard compared with the heat of combustion of the tested sample by burning 1.00 ± 0.05 g of the sample in the bomb calorimeter. The heat of combustion was computed based on the temperatures before and after combustion.

For identification of chemical constituents in bio-oil, gas chromatography-mass spectrometry (GC-MS) was performed after bio-oil toluene extraction. First, bio-oil was centrifuged to remove the solid particles and then extracted with toluene in a ratio of oil to toluene at 3:7 (v/v). GC-MS analysis (Agilent Technologies 6890N) was performed using an HP-5MS UI column (30 m × 0.25 mm × 0.25 μm). Helium was the carrier gas at the flow rate of 1 μL min^–1. An injector temperature of 300 °C, an interface temperature of 280 °C, and a detector temperature of 250 °C were applied. The amount of the injected sample was 2 μm, and the split ratio was 10/1. The operating temperature for GC-MS analysis was controlled at 50 °C (maintained for 2 min) and then increased to 290 °C (maintained for 5 min) with the heating rate of 8 °C min^–1.

Bio-oil yields were calculated by dividing the bio-oil weight (w_bio-oil, in g) by the total weight of initial dry EFB and glycerol after microwave torrefaction (w_total, in g), as shown in eq 1:⁸⁸

1

In the case of catalytic HDO bio-oil, the weight of upgraded bio-oil was recorded for the calculation of HHVs and energy ratio.The energy ratio (r) of the product was calculated based on the ratio of HHV of upgraded HDO bio-oil (HHV_HDO, MJ kg^–1) and that of raw bio-oil (HHV₀, MJ kg^–1). The energy yield (Y_E) could be calculated when Y_HDO is the mass yield of upgraded HDO bio-oil (wt %).

2

3

The carbon yield of upgraded bio-oil products was calculated from the number of moles of carbon from CHNS/O elemental analysis in the product (n_C,product) divided by the number of moles of carbon in raw EFB (n_{C,raw material}).

4

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.0c05387.

• Biomass solvothermolysis liquefaction system for production of heavy bio-oil and light bio-oil from a palm empty fruit bunch (EFB) as well as extracted light bio-oil (LBO), heavy bio-oil (HBO), and crude carbon; textural properties of reduced catalysts, i.e., BET surface area, pore size diameter, and pore volume as well as nitrogen adsorption/desorption isotherms of all calcined catalysts; gas chromatography-mass spectroscopic analysis of products from upgrading bio-oil using hydrodeoxygenation catalyzed by CoMo/Al₂O₃ and NiMo/Al₂O₃ at different temperatures (300 and 350 °C) at initial 2 MPa H₂ pressure for 1 h reaction time; comparison of main products of hydrodeoxygenated bio-oil from this study and previous studies (PDF)

The authors declare no competing financial interest.