Guerbet Reactions for Biofuel Production from ABE Fermentation Using Bifunctional Ni-MgO-Al 2 O 3 Catalysts

: To upgrade biomass-derived alcohol mixtures to biofuels under solvent-free conditions, MgO–Al 2 O 3 mixed metal oxides (MMO) decorated with Ni nanoparticles (Ni–MgO–Al 2 O 3 ) are synthesized and characterized. Based on the result, Ni nanoparticles are highly dispersed on the surface of MgAl MMO. As the Ni loading content varies from 2 to 10 wt.%, there is a slight increase in the mean Ni particle size from 6.7 to 8.5 nm. The effects of Ni loading amount, reducing temperature, and Mg/Al ratio on the conversion and product distribution are investigated. With the increase in both the Ni loading amount and reducing temperature, dehydrogenation (the ﬁrst step of the entire reaction network) is accelerated. This results in an increase in the conversion process and a higher selectivity for the dialkylated compounds. Due to the higher strength and density of basic sites under high Mg/Al ratios, double alkylation is preferred and more long-chain hydrocarbons are obtained. A conversion of 89.2% coupled with a total yield of 79.9% for C 5 –C 15 compounds is acquired by the as-prepared catalyst (prepared with Ni loading of 6 wt.%, reducing temperature of 700 ◦ C, and Mg/Al molar ratio of 3. After four runs, the conversion drops by 17.1%, and this loss in the catalytic activity can be attributed to the decrease in the surface area of the catalyst and the increase in the Ni mean particle size. or butanol may occur with the generation of hydrogen. As mentioned by Fu and Gong [43], nickel nanoparticles possess catalytic activity for the steam reforming of alcohols. The dehydrogenation of alcohols, decarbonylation of aldehydes, water–gas shift reaction, and CH 4 conversion are as follows. The activity of alcohol condensation reaction is highly related to the acidity and basicity of the Scheme 1. Catalytic upgrading mechanism of ABE mixture. catalyst, Ni loading amount: 6 wt.%, temperature used in the reduction of catalyst: 700 ◦ C, 240 ◦ C, 20


Introduction
The ever-growing concerns with regards to the continual depletion of fossil fuel reserves and the increasing severity of the environmental issues have urged the hastened development of clean energy sources. Biomass has received increasing attention as one of the most promising clean energy sources due to its abundant and renewable nature. In particular, the conversion of biomass into transportation fuels has emerged as a critical research focus in chemistry and engineering-related fields [1][2][3][4][5][6][7][8][9][10][11][12].
Step 1: Solid biomass has to be converted into platform chemicals with better catalytic activity to proceed with further treatments.
Step 2: Conversion of platform chemicals into biofuels via C-C coupling reaction and hydrodeoxygenation (HDO), when necessary [23][24][25]. Various strategies can be implemented during the first step of the conversion process, whereby these strategies can be categorized into two parts: (1) Thermochemical process can be conducted under high temperature and/or pressure, and it can later be combined with chemical upgrading, e.g., Fischer-Tropsch synthesis. (2) The hydrolysis process can be carried out so that small molecules containing oxygen functional groups, such as acetone n-butanol-ethanol (ABE) fermentation products [26], can be obtained using biological or chemical means. After 01-075-1525). The diffraction peak located at 35.1 • can be assigned to Al 2 O 3 , which overlaps with the diffraction peak of MgO. The characteristic peaks of Ni are not observed for 0 wt.% Ni loading, which is indicative that the pristine sample does not contain Ni. As the Ni loading increases to 2 wt.%, weak XRD peaks of Ni can be barely observed, which suggests that there is a low percentage of Ni in the catalyst with high dispersity. As the Ni loading increases, three distinct peaks at 44.5 • , 51.8 • , and 76.4 • can be observed, which correspond to (111), (200), and (220) planes of metallic Ni (JCPDS 03-065-0380), respectively. This XRD result suggests the successful formation of Ni nanoparticles. As shown in Table 1, with the increase in Ni loading from 2 to 10 wt.%, the average crystallite size of Ni nanoparticle increases slightly from 6.7 to 8.5 nm, based on the Scherrer equation.

Characterization of the as-Prepared Catalyst
XRD spectrums of Ni-MgO-Al2O3 catalysts with various Ni loading are presented in Figure 1. Distinct peaks located at 2θ = 36.1°, 43.1°, 62.6°, and 79.0° can be observed, which are consistent with those present in the standard XRD spectrum of MgO (JCPDS 01-075-1525). The diffraction peak located at 35.1° can be assigned to Al2O3, which overlaps with the diffraction peak of MgO. The characteristic peaks of Ni are not observed for 0 wt.% Ni loading, which is indicative that the pristine sample does not contain Ni. As the Ni loading increases to 2 wt.%, weak XRD peaks of Ni can be barely observed, which suggests that there is a low percentage of Ni in the catalyst with high dispersity. As the Ni loading increases, three distinct peaks at 44.5°, 51.8°, and 76.4° can be observed, which correspond to (111), (200), and (220) planes of metallic Ni (JCPDS 03-065-0380), respectively. This XRD result suggests the successful formation of Ni nanoparticles. As shown in Table 1, with the increase in Ni loading from 2 to 10 wt.%, the average crystallite size of Ni nanoparticle increases slightly from 6.7 to 8.5 nm, based on the Scherrer equation.  To investigate the morphology of the as-prepared catalyst, SEM, and TEM are employed. Figure S1 shows the SEM images of the as-prepared catalyst. TEM images of various catalysts prepared with different Ni loadings of 2, 6, and 8 wt.% are shown in Figure  2a-c, respectively. Ni nanoparticles are clearly observed as dark spots in the TEM images, and they are highly dispersed on the surface of MgO-Al2O3. By measuring the size of more than 150 nanoparticles observed in the TEM images, corresponding histograms of Ni particle size distributions for catalysts with 2, 6, and 8 wt.% Ni loadings can be derived, and they are presented in Figure 3a-c, respectively. The average particle size of Ni nanoparticles is estimated based on a number-weighted diameter (  To investigate the morphology of the as-prepared catalyst, SEM, and TEM are employed. Figure S1 shows the SEM images of the as-prepared catalyst. TEM images of various catalysts prepared with different Ni loadings of 2, 6, and 8 wt.% are shown in Figure 2a-c, respectively. Ni nanoparticles are clearly observed as dark spots in the TEM images, and they are highly dispersed on the surface of MgO-Al 2 O 3 . By measuring the size of more than 150 nanoparticles observed in the TEM images, corresponding histograms of Ni particle size distributions for catalysts with 2, 6, and 8 wt.% Ni loadings can be derived, and they are presented in Figure 3a-c, respectively. The average particle size of Ni nanoparticles is estimated based on a number-weighted diameter (d = ∑ n i d i /∑ n i , n i is the number of counted Ni particles with a diameter of d i ) with values of 6.8 and 7.8 nm for catalysts with 2 and 8 wt.% Ni loadings, respectively. This result confirms that the mean particle size of Ni nanoparticles increases slightly with the increase in Ni loading, which is consistent with the XRD results. Based on the high-resolution TEM image shown in Figure 2d, a lattice fringe of 0.203 nm can be clearly observed for the Ni nanoparticle, which corresponds to the (111) plane of Ni [42]. ber of counted Ni particles with a diameter of i d ) with values of 6.8 and 7.8 nm for catalysts with 2 and 8 wt.% Ni loadings, respectively. This result confirms that the mean particle size of Ni nanoparticles increases slightly with the increase in Ni loading, which is consistent with the XRD results. Based on the high-resolution TEM image shown in Figure  2d, a lattice fringe of 0.203 nm can be clearly observed for the Ni nanoparticle, which corresponds to the (111) plane of Ni [42].   ber of counted Ni particles with a diameter of i d ) with values of 6.8 and 7.8 nm for catalysts with 2 and 8 wt.% Ni loadings, respectively. This result confirms that the mean particle size of Ni nanoparticles increases slightly with the increase in Ni loading, which is consistent with the XRD results. Based on the high-resolution TEM image shown in Figure  2d, a lattice fringe of 0.203 nm can be clearly observed for the Ni nanoparticle, which corresponds to the (111) plane of Ni [42].   To provide greater details to the dispersibility of Ni nanoparticles in the as-prepared catalyst, energy-dispersive X-ray spectroscopy (EDS) elemental mapping is conducted for Ni-MgO-Al 2 O 3 catalyst with 6 wt.% Ni loading. As shown in Figure 4, the elemental distributions of Mg, Al, and Ni in the sample are highly uniform. This result suggests that Ni nanoparticles are homogeneously distributed across the well-mixed MgO-Al 2 O 3 binary oxides.
To provide greater details to the dispersibility of Ni nanoparticles in the as-prepared catalyst, energy-dispersive X-ray spectroscopy (EDS) elemental mapping is conducted for Ni-MgO-Al2O3 catalyst with 6 wt.% Ni loading. As shown in Figure 4, the elemental distributions of Mg, Al, and Ni in the sample are highly uniform. This result suggests that Ni nanoparticles are homogeneously distributed across the well-mixed MgO-Al2O3 binary oxides. BET surface areas, pore volumes, and pore size distributions of Ni-MgO-Al2O3 catalysts with various Ni loading are summarized in Table 1. It can be observed that all catalysts possess high surface areas (larger than 200 m 2 /g), which is vital in providing a large contact area between the catalyst and the reactant, thus contributing toward high catalytic activity. With the increase in Ni loading, various parameters such as surface area, pore volume, and mean pore diameter exhibit a decreasing trend. Note that as Ni loading increases from 0 to 10 wt.%, the surface area of the sample decreases from 267.1 to 227.2 m 2 /g. Meanwhile, the pore volume of the sample also decreases from 0.80 to 0.63 cm 3 /g. This observation may be due to the increased occupancy of Ni nanoparticles in the sample as the Ni loading increases. Pore size distributions and N2 isotherms are provided in Figure 5a,b, respectively. The pore size distribution, determined using the Barrett, Joyner, and Halenda method, illustrates that all the as-prepared catalysts contain mesopores with a mean pore size of approximately 6 nm. The N2 adsorption-desorption isotherms of all the as-prepared catalysts reveal a typical Type IV isotherm with a well-defined N2 hysteresis loop at relative pressures of 0.7-1.0. As such, based on this result, the as-prepared catalysts should possess mesoporous structures.  BET surface areas, pore volumes, and pore size distributions of Ni-MgO-Al 2 O 3 catalysts with various Ni loading are summarized in Table 1. It can be observed that all catalysts possess high surface areas (larger than 200 m 2 /g), which is vital in providing a large contact area between the catalyst and the reactant, thus contributing toward high catalytic activity. With the increase in Ni loading, various parameters such as surface area, pore volume, and mean pore diameter exhibit a decreasing trend. Note that as Ni loading increases from 0 to 10 wt.%, the surface area of the sample decreases from 267.1 to 227.2 m 2 /g. Meanwhile, the pore volume of the sample also decreases from 0.80 to 0.63 cm 3 /g. This observation may be due to the increased occupancy of Ni nanoparticles in the sample as the Ni loading increases. Pore size distributions and N 2 isotherms are provided in Figure  5a,b, respectively. The pore size distribution, determined using the Barrett, Joyner, and Halenda method, illustrates that all the as-prepared catalysts contain mesopores with a mean pore size of approximately 6 nm. The N 2 adsorption-desorption isotherms of all the as-prepared catalysts reveal a typical Type IV isotherm with a well-defined N 2 hysteresis loop at relative pressures of 0.7-1.0. As such, based on this result, the as-prepared catalysts should possess mesoporous structures.
To provide greater details to the dispersibility of Ni nanoparticles in the as-prepared catalyst, energy-dispersive X-ray spectroscopy (EDS) elemental mapping is conducted for Ni-MgO-Al2O3 catalyst with 6 wt.% Ni loading. As shown in Figure 4, the elemental distributions of Mg, Al, and Ni in the sample are highly uniform. This result suggests that Ni nanoparticles are homogeneously distributed across the well-mixed MgO-Al2O3 binary oxides. BET surface areas, pore volumes, and pore size distributions of Ni-MgO-Al2O3 catalysts with various Ni loading are summarized in Table 1. It can be observed that all catalysts possess high surface areas (larger than 200 m 2 /g), which is vital in providing a large contact area between the catalyst and the reactant, thus contributing toward high catalytic activity. With the increase in Ni loading, various parameters such as surface area, pore volume, and mean pore diameter exhibit a decreasing trend. Note that as Ni loading increases from 0 to 10 wt.%, the surface area of the sample decreases from 267.1 to 227.2 m 2 /g. Meanwhile, the pore volume of the sample also decreases from 0.80 to 0.63 cm 3 /g. This observation may be due to the increased occupancy of Ni nanoparticles in the sample as the Ni loading increases. Pore size distributions and N2 isotherms are provided in Figure 5a,b, respectively. The pore size distribution, determined using the Barrett, Joyner, and Halenda method, illustrates that all the as-prepared catalysts contain mesopores with a mean pore size of approximately 6 nm. The N2 adsorption-desorption isotherms of all the as-prepared catalysts reveal a typical Type IV isotherm with a well-defined N2 hysteresis loop at relative pressures of 0.7-1.0. As such, based on this result, the as-prepared catalysts should possess mesoporous structures.

Catalytic Upgrading of ABE Mixture
Scheme 1 shows the illustration of the catalytic upgrading mechanism of ABE mixtures to long-chain compounds. As illustrated in Scheme 1, three main reactions are involved during the catalytic upgrading process. Part A: Alkylation reactions producing ketones. Part B: Guerbet reactions generating alcohols with longer chains. Part C: Selfcondensation of acetones through aldol condensation. Although the detailed mechanism of the Guerbet reaction is still controversial, it is generally believed that both the Guerbet reaction and alkylation reaction involve a series of processes, such as dehydrogenation of alcohol, aldol condensation between ketone (or aldehyde) and aldehyde, dehydration, and hydrogenation.

Catalytic Upgrading of ABE Mixture
Scheme 1 shows the illustration of the catalytic upgrading mechanism of ABE mixtures to long-chain compounds. As illustrated in Scheme 1, three main reactions are involved during the catalytic upgrading process. Part A: Alkylation reactions producing ketones. Part B: Guerbet reactions generating alcohols with longer chains. Part C: Selfcondensation of acetones through aldol condensation. Although the detailed mechanism of the Guerbet reaction is still controversial, it is generally believed that both the Guerbet reaction and alkylation reaction involve a series of processes, such as dehydrogenation of alcohol, aldol condensation between ketone (or aldehyde) and aldehyde, dehydration, and hydrogenation. Scheme 1. Catalytic upgrading mechanism of ABE mixture.
It is commonly acknowledged that nanoparticles can facilitate the dehydrogenation/hydrogenation process. As such, loading amount and dispersibility of nanoparticles can play pivotal roles in influencing the catalytic activity of the catalyst. The growth of carbon chain molecules can be produced by aldol condensation. For example, aldol condensation between acetone and acetaldehyde yields 2-pentanone (2-C5), and with further aldehyde condensation between 2-C5 and acetaldehyde, 4-heptaneone (4-C7) can be obtained. Part D: Alcohols (C5-OH, C7-OH, C9-OH, and C11-OH) can be generated via the hydrogenation of corresponding ketones as shown in Scheme 1. Ketone hydrogenation requires additional hydrogen resources, which may come from two ways. First, it may be released during the aldoesterification process (Scheme 2). However, neither ethyl acetate nor butyl butyrate is observed in our reaction system, and therefore these products can be ruled out. Second, steam reforming of ethanol or butanol may occur with the generation of hydrogen. As mentioned by Fu and Gong [43], nickel nanoparticles possess catalytic activity for the steam reforming of alcohols. The dehydrogenation of alcohols, decarbonylation of aldehydes, water-gas shift reaction, and CH4 conversion are as follows. The activity of alcohol condensation reaction is highly related to the acidity and basicity of the Scheme 1. Catalytic upgrading mechanism of ABE mixture.
It is commonly acknowledged that nanoparticles can facilitate the dehydrogenation/hydrogenation process. As such, loading amount and dispersibility of nanoparticles can play pivotal roles in influencing the catalytic activity of the catalyst. The growth of carbon chain molecules can be produced by aldol condensation. For example, aldol condensation between acetone and acetaldehyde yields 2-pentanone (2-C 5 ), and with further aldehyde condensation between 2-C 5 and acetaldehyde, 4-heptaneone (4-C 7 ) can be obtained. Part D: Alcohols (C5-OH, C7-OH, C9-OH, and C11-OH) can be generated via the hydrogenation of corresponding ketones as shown in Scheme 1. Ketone hydrogenation requires additional hydrogen resources, which may come from two ways. First, it may be released during the aldoesterification process (Scheme 2). However, neither ethyl acetate nor butyl butyrate is observed in our reaction system, and therefore these products can be ruled out. Second, steam reforming of ethanol or butanol may occur with the generation of hydrogen. As mentioned by Fu and Gong [43], nickel nanoparticles possess catalytic activity for the steam reforming of alcohols. The dehydrogenation of alcohols, decarbonylation of aldehydes, water-gas shift reaction, and CH 4 conversion are as follows. The activity of alcohol condensation reaction is highly related to the acidity and basicity of the catalyst, which indicates that the optimization of acid/base strength or acid/base amount of catalyst can play a significant role. According to the abovementioned analyses, the effects of Ni amount on the ABE conversion yield and product distribution are firstly investigated. As shown in Figure 6, when ABE conversion is conducted using MMO catalyst (without Ni nanoparticle), an ABE conversion yield of 8.4% with two main products, i.e., 2-C5 and 2-C7 (monoalkylation of acetone with ethanol and butanol), are obtained. This result clearly shows that Mg-Al MMO exhibits low activity toward the dehydrogenation of ethanol and 1-butanol. Interestingly, Scheme 2. Esterification between alcohol and aldehyde.
According to the abovementioned analyses, the effects of Ni amount on the ABE conversion yield and product distribution are firstly investigated. As shown in Figure 6, when ABE conversion is conducted using MMO catalyst (without Ni nanoparticle), an ABE conversion yield of 8.4% with two main products, i.e., 2-C 5 and 2-C 7 (monoalkylation of acetone with ethanol and butanol), are obtained. This result clearly shows that Mg-Al MMO exhibits low activity toward the dehydrogenation of ethanol and 1-butanol. Interestingly, as Ni nanoparticle is incorporated into Mg-Al MMO, ABE conversion yield is significantly improved as observed in Figure 6. Note that as Ni loading in the catalyst increases from 0 to 2 wt.%, ABE conversion yield increases drastically from 8.4% to 58.8%. After which, as Ni loading increases from 2 wt.% to 6 wt.%, ABE conversion yield continues to increase steadily from 58.8% to 89.2%. When the nickel loading is more than 6 wt.%, the yield decreases from 89.2% to 86.48% with increasing nickel loadings. Furthermore, it can be observed that with the increase in Ni loading, the main product changes from mono-alkylated compounds (C5 and 2-C7) to double-alkylated ones with longer carbon chains (4-C7 to C15). For instance, the total yield of double-alkylated compounds reaches 79.88% when the Ni loading is 6 wt.%. Two key reasons can be used to explain such phenomenon: (1) As Ni nanoparticle exhibits high dehydrogenation activity, increasing Ni loading would translate to the production of more aldehydes, which can then act as reactants for subsequent aldol condensation. This process can lead to a significant improvement in double alkylation. However, over-high Ni loading content is not valuable for the conversion yield. (2) C=C bonds in α, β-unsaturated ketones are kinetically and thermodynamically favored by the Ni site, and therefore saturated ketones are generated [44]. According to the abovementioned analyses, the effects of Ni amount on the ABE conversion yield and product distribution are firstly investigated. As shown in Figure 6, when ABE conversion is conducted using MMO catalyst (without Ni nanoparticle), an ABE conversion yield of 8.4% with two main products, i.e., 2-C5 and 2-C7 (monoalkylation of acetone with ethanol and butanol), are obtained. This result clearly shows that Mg-Al MMO exhibits low activity toward the dehydrogenation of ethanol and 1-butanol. Interestingly, as Ni nanoparticle is incorporated into Mg-Al MMO, ABE conversion yield is significantly improved as observed in Figure 6. Note that as Ni loading in the catalyst increases from 0 to 2 wt.%, ABE conversion yield increases drastically from 8.4% to 58.8%. After which, as Ni loading increases from 2 wt.% to 6 wt.%, ABE conversion yield continues to increase steadily from 58.8% to 89.2%. When the nickel loading is more than 6 wt.%, the yield decreases from 89.2% to 86.48% with increasing nickel loadings. Furthermore, it can be observed that with the increase in Ni loading, the main product changes from mono-alkylated compounds (C5 and 2-C7) to double-alkylated ones with longer carbon chains (4-C7 to C15). For instance, the total yield of double-alkylated compounds reaches 79.88% when the Ni loading is 6 wt.%. Two key reasons can be used to explain such phenomenon: (1) As Ni nanoparticle exhibits high dehydrogenation activity, increasing Ni loading would translate to the production of more aldehydes, which can then act as reactants for subsequent aldol condensation. This process can lead to a significant improvement in double alkylation. However, over-high Ni loading content is not valuable for the conversion yield.
(2) C=C bonds in α, β-unsaturated ketones are kinetically and thermodynamically favored by the Ni site, and therefore saturated ketones are generated [44]. Other than the Ni loading contents, the temperature used in the reduction of the catalyst can also influence the catalytic activity of the catalyst. As such, various temperatures are used in the preparation process, and the corresponding catalytic performances of the as-prepared catalysts are shown in Figure 7. It is clearly shown that 7% ABE conversation yield with C5 as the sole product is achieved when catalysts that are reduced at 400 °C and 500 °C are used. As the temperature used in the reduction of catalyst increases to 600 Other than the Ni loading contents, the temperature used in the reduction of the catalyst can also influence the catalytic activity of the catalyst. As such, various temperatures are used in the preparation process, and the corresponding catalytic performances of the as-prepared catalysts are shown in Figure 7. It is clearly shown that 7% ABE conversation yield with C5 as the sole product is achieved when catalysts that are reduced at 400 • C and 500 • C are used. As the temperature used in the reduction of catalyst increases to 600 • C, an increase in the ABE conversion yield is observed, with long-chain ketones and alcohols as the products. As the temperature increases from 600 • C to 800 • C, a total yield that increases from 68.4% to 88.6% is recorded. This observation is largely attributed to the fact that HT precursors would not be able to completely convert to MMO at a temperature lower than 500 • C, which results in higher catalytic activity for subsequent aldol condensation [45,46]. On the other hand, Ni nanoparticle is expected to catalyze the dehydrogenation of alcohols, while Ni 2+ shows low dehydrogenation activity.
To verify the abovementioned hypothesis, XPS is used to further characterize the catalysts reduced at various temperatures. Binding energy values of metallic Ni are 852.7 eV (Ni 2p 3/2 ) and 870.5 eV (Ni 2p 1/2 ), while those of NiO are 854.0 eV (Ni 2p 3/2 ) and 872.5 eV (Ni 2p 1/2 ) [46,47]. As shown in Figure 8, as the temperature used in the reduction of the catalyst decreases to the range of 400 to 500 • C, a peak near 854.6 eV is observed, which indicates that Ni species exists primarily in the form of NiO. On the other hand, as the temperature increases from 600 to 800 • C, a peak around 852.6 eV can be clearly observed. The shift in the binding energy toward lower values may be attributed to the change in the configuration of Ni in the MgO-Al 2 O 3 matrix. As indicated in Table 2, the amount of Ni nanoparticles in the catalyst increases from 7% to 87% as the temperature used in the reduction of catalyst increases from 400 to 800 • C. °C, an increase in the ABE conversion yield is observed, with long-chain ketones and alcohols as the products. As the temperature increases from 600 °C to 800 °C, a total yield that increases from 68.4% to 88.6% is recorded. This observation is largely attributed to the fact that HT precursors would not be able to completely convert to MMO at a temperature lower than 500 °C, which results in higher catalytic activity for subsequent aldol condensation [45,46]. On the other hand, Ni nanoparticle is expected to catalyze the dehydrogenation of alcohols, while Ni 2+ shows low dehydrogenation activity. To verify the abovementioned hypothesis, XPS is used to further characterize the catalysts reduced at various temperatures. Binding energy values of metallic Ni are 852.7 eV (Ni 2p3/2) and 870.5 eV (Ni 2p1/2), while those of NiO are 854.0 eV (Ni 2p3/2) and 872.5 eV (Ni 2p1/2) [46,47]. As shown in Figure 8, as the temperature used in the reduction of the catalyst decreases to the range of 400 to 500 °C, a peak near 854.6 eV is observed, which indicates that Ni species exists primarily in the form of NiO. On the other hand, as the temperature increases from 600 to 800 °C, a peak around 852.6 eV can be clearly observed. The shift in the binding energy toward lower values may be attributed to the change in the configuration of Ni in the MgO-Al2O3 matrix. As indicated in Table 2, the amount of Ni nanoparticles in the catalyst increases from 7% to 87% as the temperature used in the reduction of catalyst increases from 400 to 800 °C.  condensation [45,46]. On the other hand, Ni nanoparticle is expected to catalyze the dehydrogenation of alcohols, while Ni 2+ shows low dehydrogenation activity. To verify the abovementioned hypothesis, XPS is used to further characterize the catalysts reduced at various temperatures. Binding energy values of metallic Ni are 852.7 eV (Ni 2p3/2) and 870.5 eV (Ni 2p1/2), while those of NiO are 854.0 eV (Ni 2p3/2) and 872.5 eV (Ni 2p1/2) [46,47]. As shown in Figure 8, as the temperature used in the reduction of the catalyst decreases to the range of 400 to 500 °C, a peak near 854.6 eV is observed, which indicates that Ni species exists primarily in the form of NiO. On the other hand, as the temperature increases from 600 to 800 °C, a peak around 852.6 eV can be clearly observed. The shift in the binding energy toward lower values may be attributed to the change in the configuration of Ni in the MgO-Al2O3 matrix. As indicated in Table 2, the amount of Ni nanoparticles in the catalyst increases from 7% to 87% as the temperature used in the reduction of catalyst increases from 400 to 800 °C.   It is well accepted that the acid-base properties of the catalyst can play a key role in influencing the aldol condensation activity. The weak Brønsted basic sites of MgAl-MMO are related to the residual surface hydroxyl groups after activation, the moderate strength Lewis sites are related to Mg 2− O 2− and Al 3+ O 2− acid-base pairs, and the strong Lewis base sites are due to the existence of low coordinated O 2 species. The lower Al dopant content and the higher Mg content led to the increase in the basic center density, which is due to the formation of coordinated unsaturated oxygen sites. Materials with high Al contents are beneficial to the dehydration of alcohols rather than dehydrogenation and condensation. Ref. [48] by varying the Mg/Al-ratios, which changes the number and strength of the acid-base sites, the selectivity can be optimized towards dehydrogenation, aldolization, and hydride-shifts. Thus, the effect of the Mg/Al ratio on the catalytic performance of the catalyst is investigated. As shown in Figure 9, as the Mg/Al ratio increases from 1 to 9, ABE conversion yield remains constant at 88.4%. However, significant changes in the product distribution are observed across the varying Mg/Al ratio. Note that as the Mg/Al ratio increases from 1 to 9, the selectivity for C 5 and 2-C 7 decreases and more C 8 -C 15 are obtained. This result indicates that double alkylation is preferable at higher Mg/Al ratios. Figure S6 shows the results of the product distribution for the catalytic coupling of the ABE mixture. 800 87 13 6.69 It is well accepted that the acid-base properties of the catalyst can play a key role in influencing the aldol condensation activity. The weak Brønsted basic sites of MgAl-MMO are related to the residual surface hydroxyl groups after activation, the moderate strength Lewis sites are related to Mg 2− O 2− and Al 3+ O 2− acid-base pairs, and the strong Lewis base sites are due to the existence of low coordinated O2 species. The lower Al dopant content and the higher Mg content led to the increase in the basic center density, which is due to the formation of coordinated unsaturated oxygen sites. Materials with high Al contents are beneficial to the dehydration of alcohols rather than dehydrogenation and condensation. Ref. [48] by varying the Mg/Al-ratios, which changes the number and strength of the acid-base sites, the selectivity can be optimized towards dehydrogenation, aldolization, and hydride-shifts. Thus, the effect of the Mg/Al ratio on the catalytic performance of the catalyst is investigated. As shown in Figure 9, as the Mg/Al ratio increases from 1 to 9, ABE conversion yield remains constant at 88.4%. However, significant changes in the product distribution are observed across the varying Mg/Al ratio. Note that as the Mg/Al ratio increases from 1 to 9, the selectivity for C5 and 2-C7 decreases and more C8-C15 are obtained. This result indicates that double alkylation is preferable at higher Mg/Al ratios. Figure S6 shows the results of the product distribution for the catalytic coupling of the ABE mixture. The acid and base properties of catalysts with various Mg/Al ratios are investigated using NH3-TPD and CO2-TPD, respectively. As shown in Figure 10a, catalysts with various Mg/Al ratios exhibit a similar profile with observable broad peaks at around 120 °C, which indicates that Ni-MgO-Al2O3 catalysts only contain weak acidic sites. As shown in Figure 10b, CO2-TPD profiles are composed of two overlapping desorption peaks centered around 150 °C (peak I) and 260 °C (peak II), which correspond to weak and moderate basic sites, respectively. The acid and base properties of catalysts with various Mg/Al ratios are investigated using NH 3 -TPD and CO 2 -TPD, respectively. As shown in Figure 10a, catalysts with various Mg/Al ratios exhibit a similar profile with observable broad peaks at around 120 • C, which indicates that Ni-MgO-Al 2 O 3 catalysts only contain weak acidic sites. As shown in Figure 10b, CO 2 -TPD profiles are composed of two overlapping desorption peaks centered around 150 • C (peak I) and 260 • C (peak II), which correspond to weak and moderate basic sites, respectively. The concentrations of the acidic sites of the catalysts with various Mg/Al ratios are listed in Table 3. The concentration of acidic sites gradually decreases to a minimum value of 0.27 µmol/g with an increase in Mg/Al ratio. On the other hand, with the increase in Mg/Al ratio, the density of weak basic site decreases from 0.69 to 0.61 µmol/g, while densities of moderate basic site and the total basic site gradually increase. The weak acidic sites in the catalyst are beneficial toward the dehydration of unstable aldol products. However, the presence of strong acid sites could potentially result in side reactions such as dehydration of alcohols to olefins. The basicity and number of basic sites play an important role in determining the product distribution. With the consideration of the results presented in Figure 9, it is suggested that there is a preferential double alkylation of ace- The concentrations of the acidic sites of the catalysts with various Mg/Al ratios are listed in Table 3. The concentration of acidic sites gradually decreases to a minimum value of 0.27 µmol/g with an increase in Mg/Al ratio. On the other hand, with the increase in Mg/Al ratio, the density of weak basic site decreases from 0.69 to 0.61 µmol/g, while densities of moderate basic site and the total basic site gradually increase. The weak acidic sites in the catalyst are beneficial toward the dehydration of unstable aldol products.
However, the presence of strong acid sites could potentially result in side reactions such as dehydration of alcohols to olefins. The basicity and number of basic sites play an important role in determining the product distribution. With the consideration of the results presented in Figure 9, it is suggested that there is a preferential double alkylation of acetone when using catalysts with higher basicity and more basic sites. This may explain the need for large amounts of alkali such as K 3 PO 4 , KOH, and K 2 CO 3 in other works [27,28].

Regeneration Performance
The reusability of the as-prepared catalyst is studied, and the result is shown in Figure 11. It can be observed that ABE conversion yield decreases from 89.2% to 72.1% after four runs. To provide some insights into the decrease in the ABE conversion yield with runs, the spent catalysts are investigated with XRD, BET, and TEM. As shown in Figure S2, there is no distinct difference between the XRD spectrums of the fresh and spent catalysts, which indicates that the crystal structure and phase of the catalyst remain unchanged after the operation. The pore size distribution and N 2 adsorption-desorption isotherms of the spent catalysts are depicted in Figure S3, with the rest of the BET results listed in Table S1. With the increase in the number of recycling runs, the specific surface area and pore volume of the spent catalyst decrease gradually. Such a result may be the key factor towards the observed catalytic activity loss. TEM image of the spent catalyst is shown in Figure S4a, and the histogram of the size distribution of Ni nanoparticles is shown in Figure S4b. The mean diameter of Ni nanoparticles increases to 8.1 nm after several cycles, which is another factor that poses a detrimental effect on the catalytic performance of the catalyst. The base density of the used catalyst decreased significantly from 1.24 to 0 µmol/g, which may be caused by the formation of MgCO 3 . The main peaks located at 400 • C and 540 • C can be attributed to the release of CO 2 from the decomposition of MgCO 3 . The decrease in the base density and surface area of the activated catalyst may lead to a decrease in its catalytic activity. several cycles, which is another factor that poses a detrimental effect on the catalytic performance of the catalyst. The base density of the used catalyst decreased significantly from 1.24 to 0 µmol/g, which may be caused by the formation of MgCO3. The main peaks located at 400 °C and 540 °C can be attributed to the release of CO2 from the decomposition of MgCO3. The decrease in the base density and surface area of the activated catalyst may lead to a decrease in its catalytic activity.    China). Deionized water was used in all reactions. All chemicals were used as received, without further purification.

Preparation of the Catalyst
Ni-MgO-Al 2 O 3 mixed metal oxide (MMO) was prepared via co-precipitation. In a typical preparation process, an aqueous solution (0.1 L) of Na 2 CO 3 (0.2 mol, 2.12 g) solution was added into an aqueous solution (0.1 L) containing Mg(NO 3 ) 2 ·6H 2 O (1.05 mol, 26.92 g), Al(NO 3 ) 3 ·9H 2 O(0.35 mol, 13.13 g), and Ni(NO 3 ) 2 ·6H 2 O(0.09 mol, 2.62 g), and the mixture was mixed via vigorous stirring. The pH of the mixture was carefully maintained in the range of 9 to 10 using an aqueous 0.3 M NaOH solution. After continuously stirring and aging overnight, the precipitate (Mg-Al MMO) was filtered and washed with deionized water until the pH of the precipitate reached 7, and the washed precipitate was dried at 100 • C overnight. The as-prepared catalyst was subsequently dried at 80 • C overnight. Finally, the catalyst was reduced at 700 • C in the presence of H 2 .

Characterization Techniques
XRD spectrum of the sample was recorded using a Shimadzu XRD-6000 diffractometer (Shimadzu, Tokyo, Japan) with Cu Kα radiation (λ = 1.5418 Å). A 2-theta value range of 5 • ≤ 2θ ≤ 90 • was used in the XRD measurement. XPS spectrum of the sample was measured using a Thermo VGESCALAB 250 spectrometer (Thermo, Waltham, MA, USA), with Mg Kα (1253.6 eV) radiation as the X-ray source. All binding energies were calibrated with reference to the position of C1s peak at 284.6 eV. The morphology of the sample was observed under a ZEISS SUPRA55 SEM (ZEISS, Jena, Germany), with an accelerating voltage of 2.0 kV. The structure, size, and lattice fringes of the sample were examined under a JEOL JEM-2011 TEM (JEOL, Tokyo, Japan), equipped with an energy-dispersive X-ray spectrometer. The specific surface area of the sample was measured (Bruker, Karlsruhe, Germany) according to the Brunauer-Emmett-Teller (BET) method based on N 2 adsorption isotherm. All samples were degassed at 180 • C for 4 h prior to the BET measurement. The acid-base properties of the catalyst were determined using NH 3 -TPD and CO 2 -TPD, respectively, which are both equipped with a thermal conductivity detector (TCD, Bruker, Karlsruhe, Germany). 0.2 g sample was pretreated in a U-tube glass under a 25 mL/min He flow at 350 • C for 1 h, and it was then cooled to 100 • C. After completing the degassing of the sample, the adsorption gas was switched to CO 2 or NH 3 , which was used to flush the U-tube glass for 30 min. After which, the temperature was increased from 100 to 850 • C at a heating rate of 10 • C/min, under a pure He atmosphere, to record the TCD signals.

Catalytic Conversion of Feedstock to Biofuel
A mixture of acetone, n-butanol, and ethanol with a molar ratio of 2.3:3.7:1 was used as a model of ABE fermentation. The catalytic conversion was conducted in a high-pressure reaction vessel, which was equipped with a magnetic stirring bar (IKA, Köln, Germany). In a typical reaction process, 1.5 g of the as-prepared catalyst and 15 g ABE mixture were added into the reactor, and the reactor was then heated to 240 • C for 20 h under a stirring speed of 800 rpm. After which, the reactor was cooled to room temperature by immersing it into an ice-water bath. The reaction product is analyzed using Shimadzu GC-2014 Chromatograph (Shimadzu, Tokyo, Japan) and Agilent GC-MS (Agilent, Palo Alto, Santa Clara, CA, USA) with DB-5 column, according to our previous work [37].

Conclusions
In summary, the potential of Ni-MgO-Al 2 O 3 as a heterogeneous catalyst in biofuel production is investigated in this work. The as-prepared catalyst shows high efficiency in upgrading ABE mixture into long-chain (C 5 -C 15 ) ketones and alcohols, which are important biofuel precursors. With the increase in Ni loading from 2 to 10 wt.%, the specific surface area of the catalyst decreases from 256.1 to 227.2 m 2 /g, while the mean diameter of Ni nanoparticles increases from 6.7 to 8.5 nm. The acid-base properties of the as-prepared catalyst can be controlled by adjusting the Mg/Al molar ratio. Based on the result, catalysts with Mg/Al molar ratio in the range of 1 to 9 all show weak acidic sites, with a decreasing concentration of these acidic sites from 0.41 to 0.27 µmol/g. In contrast, as the Mg/Al molar ratio increases from 1 to 9 and the concentration of basic site increases from 1.09 to 1.45 µmol/g, with more moderate basic sites being generated. Furthermore, it is shown that higher conversion and greater preferential for double alkylation can be realized with higher Ni loading and higher temperature used in the reduction of catalyst. The Mg/Al molar ratio has little effect on the conversion yield, but it plays a significant role in influencing the product distribution. When employing a catalyst with a high amount of Mg, significant enhancement in the selectivity for 4-C 7 to C 15 hydrocarbons is observed. The catalyst, prepared with Mg/Al ratio of 3 and 6 wt.% Ni loading, and reduced at 700 • C, can achieve a conversion yield of 89.2% with the total C5-C15 compounds yield of 79.9%. The cyclic performance of the catalyst is also investigated, whereby there is a 17.1% decrease in the conversion yield after 4 runs. This loss in the activity may be a result of the decrease in the surface area and increase in the mean Ni particle size.

Data Availability Statement:
The data presented in this paper are from published sources.