The Influence of Incorporating Alkaline Metals on Alumina for Xylitol Hydrogenolysis over Ni Catalysts Authors: Vuyolwethu O. Ndabankulu, Alisa Govender, Holger B. Friedrich, Mzamo L. Shozi Published: 2026 ABSTRACT Bifunctional nickel catalysts, Ni/MO-Al2O3 (M = Ba, Ca, Mg, and La) were prepared by the coprecipitation method for the support, followed by wet impregnation of Ni, and used for the hydrogenolysis of xylitol to glycols. According to the XRD data, the crystal structure of the support was maintained after the loading of Ni. All the prepared catalysts were mesoporous, according to N2 physisorption isotherms. XPS revealed Ni 2p peaks, which were assigned to NiO on the surfaces of the catalysts. The bifunctional catalysts showed good xylitol conversions and glycol selectivity over a period of 24 h. The surface structure suggests that Ni/BaO-Al2O3 (Ni/BA) promotes dehydrogenation and the retro-aldol reaction, which favors C2–C3 bond cleavage. The incorporation of alkaline metals reduced the formation of cracked products, such as methane and carbon dioxide, while increasing glycol yields. A combined yield of 80% of propylene glycol (PG) and ethylene glycol (EG) was thus obtained over Ni/BA at 200 °C and 40 bar H2 pressure. Introduction The development of environmentally and eco-friendly methods to produce chemicals and fuels is one of the best ways to mitigate current environmental problems, such as an increase in global air pollution and climate change. Biomass offers renewable and sustainable raw materials that can be used as feedstock to produce many liquid chemicals and biofuels. Research has shown that upgrading lignin and cellulosic biomass provides an opportunity to develop renewable resources to produce chemicals and fuels from biomass materials and plants that do not compete with food production. The major challenge facing the upgrading of biomass-derived compounds, such as polyols, is the high oxygen content. The complex mixture of oxygenates has fewer applications, such as in liquid chemicals and fuels. Hydrogenation, hydrogenolysis, and dehydration processes have been employed to upgrade biomass feedstocks. − Xylitol, which is derived from xylose, is one of the platform sugar alcohols that can be upgraded to form various valuable chemicals, such as glycols and alcohols. Hydrogenolysis of sugar alcohols, such as glycerol, xylitol, and sorbitol, is achieved using metals like Pt, Ru, Pd, Ni, and Cu. − The initial step is the dehydrogenation of the sugar alcohol, which is done by the active metal, followed by a retro-aldol reaction, which is base-facilitated, and then leads to C–C cleavage. Studies on the hydrogenolysis of sugar alcohols, such as glycerol, have shown that in base-facilitated mechanisms, glycerol dehydrogenation to form glyceraldehyde, which is then hydrated to produce 2-hydroxyacrolein. This then undergoes keto–enol tautomerism to form hydroxyacetone. This mechanism is similar to that occurring in the hydrogenolysis of xylitol. Liu et al. studied the efficiency of using a NiCu-SiO2 bimetallic catalyst for hydrogenolysis of xylitol using Ca­(OH)2 as the base. They obtained up to 94.4% xylitol conversion using 10Ni80Cu-SiO2 and Ca­(OH)2 at 473 K under 6 MPa H2 with 31.6% 1,2-propylene glycol (PG) selectivity and 28.8% ethylene glycol (EG) selectivity. They reported that xylitol conversion and glycol selectivity are dependent on the Ni/Cu ratio and the method used to prepare the catalyst. The coprecipitation method was reported to be the best method. They also reported that retro-aldol intermediates, glyceraldehyde and glycolaldehyde, can be transformed into glycolic acid and lactic acid by Ca­(OH)2. It has been reported that the addition of a base as a cocatalyst in hydrogenolysis is required to facilitate the C–C bond cleavage. However, it has also been reported that the addition of bases such as Ca­(OH)2, Ba­(OH)2, NaOH, and La­(OH)3. leads to the formation of lactic acid and degradation of glycols to alcohols (methanol, ethanol, propanol) and gas products such as methane and carbon dioxide, and requires extensive separation. , To address these issues, bifunctional catalysts with basic properties have been investigated. Jin et al. studied hybrid Cu/CaO-Al2O3 prepared by the coprecipitation method in hydrogenolysis of xylitol (2.7% solution) at 230 °C and 7.6 MPa H2 pressure, and no additional base was added. The authors reported 98.1% conversion after 6 h with 46.1% PG and 15.4% EG selectivity; they also observed a reduction in lactic acid formation. In this work, we report the influence of incorporating alkaline earth metals in alumina and impregnating nickel on the hydrogenolysis of xylitol. This was done to modify the basic sites, which are vital in xylitol conversion to form glycols with reduced amounts of alcohol and gas product formation. Nickel was chosen because of its excellent hydrogen molecule activation as well as good C–C bond cleavage over copper metal. This was expected to improve ethylene and propylene glycol selectivity and minimize the C–O cleavage. Results and Discussion Catalyst Characterization Powder X-Ray Diffraction (XRD) The crystallinity and phases of the prepared materials were investigated using XRD in the 2θ range between 5° and 90° (Figure S1). For MgO-Al2O3 (MA), one sees that it exists as a mixed phase, which includes the MgAl2O4 spinel phase. This is supported by the presence of (111), (220), (311), (400), and (422) lattice planes. , They are found at 2θ values, 19.4°, 31.7°, 36.3°, 45°, and 65° (PDF#21–1152). As shown in Figure S1a, MA spinel had higher crystallinity compared to the rest of the supports. BA also exhibited spinel peaks; however, the peaks were less intense and broader than those of MA. This was attributed to the calcination temperature not being high enough to form the spinel. The presence of Al2O3 was evident in the diffraction patterns for BA and CA supports. CA also exhibited a CaO lattice plane at 65°. LA was found to be amorphous, with a broad peak at 27° attributed to La2O3. Upon loading 15 wt% Ni on the prepared supports, additional diffraction peaks were observed, which were due to the presence of NiO (JCPDS, No. 04–0835). Electron Microscopy A high-resolution transmission electron microscope (HRTEM) was used to obtain elemental mapping and selected area electron diffraction (SAED) of the prepared catalysts. Figure S2a shows EDS mapping for each catalyst. The EDS mapping showed that there is an even distribution of the metal within the catalysts, which was attributed to the choice of catalyst preparation, as coprecipitation tends to offer good distribution. SAED, in Figure S2b, agreed with the data obtained from XRD, and the NiO lattice planes were observed, but the catalysts had poor crystallinity. In contrast, Ni/CaO-Al2O3 (Ni/CA) showed sharp NiO diffraction peaks in the XRD, which were confirmed by bright rings in each SAED ring corresponding to the NiO lattice plane. Ni/La2O3–Al2O3 (Ni/LA) showed poor crystallinity in XRD, which was supported by SAED, since it showed clustered diffraction spots. TEM images could not be used to distinguish between the Ni particles and the support. N2-Physisorption The obtained isotherms from N2-physisorption analysis of the supports are presented in Figure , as are their pore size distribution curves. All of the synthesized supports had Type IV isotherms and Type H2 Hysteresis loops, which correspond to mesoporous materials according to the IUPAC classification. The textural properties of the supports are shown in Table S1. MA showed the highest surface area of all of the supports, followed by BA, then CA, and LA, with the lowest surface area. Pore size distribution curves were plotted using the desorption isotherm and applying the Barrett–Joyner–Halenda (BJH) method. All of the mixed metal oxides displayed a similar profile range. MA had the largest average pore size, followed by BA and CA, with LA having the lowest average pore size distribution. After impregnating Ni, the shape of the isotherms remained the same for all of the catalysts. However, there was a drop in the BET surface area, as seen in Table . This was attributed to Ni filling the pores of the support, which reduced the amount of N2 adsorbed by the catalyst. Ni/MA had the largest surface area of 163 m2/g. There was an increase in the pore size of Ni/MA (3.8 nm) compared to MA (3.6 nm). This increase is accompanied by a drop in the pore volume from 0.27 cm3/g (MA) to 0.19 cm3/g (Ni/MA). There is also a drop in surface area after adding Ni on the BA support. For both Ni/CA and Ni/LA, there was a minimal change in surface area after adding Ni, and this was attributed to Ni being loaded in the pores of the support instead of blocking them, as observed for Ni/MA and Ni/BA. In both catalysts, there is a decrease in pore volume and pore size, which confirms that Ni is incorporated in the pores rather than clustering on the surface of the support. All the prepared catalysts had similar pore size distribution profiles. N2-physisorption isotherms and pore size distribution curves for Ni catalysts. 1Textural Properties of the Ni CatalystssampleBET surface area (m2/g)BJH pore size (nm)BJH pore volume (cm3/g)Ni/BA753.30.05Ni/CA1153.40.11Ni/MA1633.80.19Ni/LA453.80.05 Textural Properties of the Ni Catalysts X-ray Photoelectron Spectroscopy (XPS) XPS was applied to get more information about the surface composition and chemical state of the prepared catalysts. XP spectra are shown in Figure . Looking at Ni for all of the catalysts, after resolving the spectra, Ni 2p consists of two spin–orbit doublet peaks. The Ni 2p3/2 peak was deconvoluted, and it was found that nickel species existed in different states, NiO x , and Ni-M (M = Ba, Ca, La, Mg). − NiO has its Ni 2p3/2 peak position at binding energy (BE) between 854 and 854.7 eV. This suggests that nickel strongly interacts with the support. There were no Ni­(OH)2 peaks observed in the prepared catalysts, and this is evident in both the O 1s and Ni 2p spectra. The presence of Ni3+ on the surface (856.4 eV) of the prepared catalysts can be postulated to arise from Ni–metal interaction. This is supported by the existence of an O 1s peak at 530.3 eV (O2–), which corresponds to lattice oxygen, and this oxygen peak is associated with NiO x . The peak at 533 eV is assigned to adsorbed oxygen species of hydroxyl groups (OH), and this can be attributed to adsorbed water and/or carbonates. Oxygen vacancies were also detected at a BE of 532.5 eV (O1), and these were due to the supports. Looking at Al 2p, it has a single spin–orbit doublet. It is observed that it exhibits the same profile for all of the prepared catalysts. The broad peak at 74.4 eV is attributed to Al2O3. , The Ba 3d5/2 spectrum for the Ni/BA catalyst was deconvoluted to two peaks. The lower BE of 779.5 eV is assigned to BaO, and BE 782 eV is assigned to Ba bridging with two oxygen atoms. − For the Ni/MA catalyst, the Mg 1s spectrum showed a peak at 1304.5 eV corresponding to MgO. Lanthanum in Ni/LA exists only in the La3+ state, and this was noted by the existence of a La d5/2 peak. The peak appearing at 834.5 eV corresponded to La2O3, with a satellite peak at 839.5 eV. La2O3 usually has an O 1s peak with a BE of 528.6 eV, and this peak was shifted to 530 eV in the case of Ni/LA. The shift in the BE may be due to the formation of La2CO3 on the surface, as it has been reported that La2O3 can absorb moisture and CO2 from the atmosphere to form La­(OH)3 and La2CO3. The Ni/CA spectra showed two main peaks, which are characteristic of Ca 2p with BE at 346.5 and 351.5 eV. These peaks were assigned to CaO. There was no evidence of CaCO3. Ba 3d, Ca 2p, La 3d, and Mg 1s XPS spectra of the prepared supports (a–d) and Ni 2p XPS spectra of the catalysts (e–h). H2–Temperature-Programmed Reduction (TPR) The TPR data of the prepared catalysts, shown in Figure , reveal that NiO in all four catalysts was reducible. However, all catalysts had different profiles, which was attributed to the difference in the metal–support interaction. The Ni/BA catalyst showed a reduction in three regions at 200 °C, which was attributed to weakly bound NiO. A small peak was also observed at 280 °C for the reduction of weakly bound NiO. The bulk of NiO on the support was reduced at 350 °C. Ni/CA had a similar reduction profile; however, the peak that appeared at 350 °C for Ni/BA was shifted to 320 °C for Ni/CA. The reduction peak at 280 °C was pronounced for Ni/CA. When looking at the profile for the Ni/MA catalyst, a broad peak between 550 and 600 °C was observed. This was assigned to Ni ions that substituted Mg ions in the support, since it is well reported that at high temperatures, NiO can form a solid solution with MgO. − Ni/LA only showed two reduction peaks, at 200 and 320 °C, and these were ascribed to weakly bound NiO and bulk NiO, respectively. The ease of reduction achieved in these catalysts correlates with the data obtained from XPS, where good Ni–metal interaction was seen, which then led to easily reduced NiO. H2–TPR profile of the prepared catalysts. CO2–Temperature-Programmed Desorption (TPD) Profiles in Figure show that each catalyst had different adsorption capacities for the dissolved CO2. Taking a closer look at Ni/LA, it is seen that it exhibited weak basic sites at 220 °C, moderate basic sites at 300 °C, and strong basic sites at 500 °C, which were attributed to O2–. Ni/CA had a similar profile, but the desorption peaks were shifted to higher temperatures, meaning it had a stronger affinity for CO2 than Ni/LA. The Ni/MA catalyst showed weak basic sites at 220 °C, moderate sites at 290 °C, and strong sites at 500 °C. Ni/BA showed only a strong desorption peak at 500 °C, ascribed to surface oxygen. The total basicity of each catalyst is shown in Table . CO2-TPD profiles of the supported Ni catalysts. 2H2 and CO2-Chemisorption Data for the Prepared Nickel CatalystscatalystNi metallic area (m2 g–1) Ni dispersion (%) Ni crystallite size (nm)total basicity (mmol g–1) 15Ni/MA3.53.8280.1715Ni/CA4.14.1240.3915Ni/LA9.49.4100.1615Ni/BA4.44.4220.05aH2-chemisoption.bCO2-TPD. H2 and CO2-Chemisorption Data for the Prepared Nickel Catalysts H2-chemisoption. CO2-TPD. Catalytic Testing Unsupported Ni was tested, and the conversions and product distribution are shown in Figure S3. It showed good xylitol conversion, from 53% after 6 h to 58%, 87%, and 100% after 12, 18, and 24 h, respectively. Although unsupported Ni showed good conversion, it gave poor glycol selectivity over the tested period. This was attributed to the lack of basic sites to facilitate the retro-aldol reaction, which produces glycols. Without the basic sites, Ni promotes dehydrogenation, which is subsequently followed by isomerization, and results in the formation of byproducts and excess methane. High selectivity toward C4 products was observed, especially erythritol and other polyols. The prepared catalysts Ni/BA, Ni/CA, Ni/LA, and Ni/MA were evaluated for the hydrogenolysis of xylitol to EG and PG at 200 °C and 40 bar. The results are shown in Figure . All the prepared bifunctional catalysts showed good conversion after a period of 24 h. Ni/BA showed the highest conversions over the period, reaching 100%, and this was followed by Ni/CA with 96%, while Ni/LA and Ni/MA had similar overall conversions after 24 h, with 79% and 81%, respectively. Ni facilitates xylitol dehydrogenation to form either ketone or aldehyde intermediates. − The incorporated alkali metal facilitates the retro-aldol reaction, which leads to C–C bond cleavage. Ni then hydrogenated the resulting intermediates. Selectivity–time profiles of xylitol conversion and product selectivity over the prepared catalysts. Reaction conditions: 15 wt% xylitol in water, 0.5 g of catalyst, Temp = 200 °C, H2-pressure = 40 bar. Alcohols: methanol, ethanol, 1-propanol, and 2-propanol. Others: C4 aldose, 1,2-butanediol, carboxylic acids, formic acid, lactic acid, and gaseous products: CH4, CO2, CO, and trace amounts of ethane and propane. After 6 h of reaction, Ni/BA showed 41% conversion with 31% and 26% selectivities toward PG and EG, respectively. Glycerol had 22% selectivity, followed by alcohols (methanol, ethanol, and propanol) with 11%, while other products came to 10%. The other products in the gas phase included CH4, CO, and CO2. In the liquid phase, these include butanol, 1,2-butanediols (1,2-BDO), C4 alditols, and trace amounts of carboxylic acids. When the reaction time increased to 12 h, the conversion increased to 51%, and this was accompanied by increases in PG and EG selectivities to 35 and 28%, respectively. This led to a decrease in glycerol selectivity to 18% and alcohol selectivity to 9%. This was expected since both PG and glycerol are formed from glyceraldehyde. A sharp increase in conversion to 82% was observed after 18 h; however, product selectivity remained the same. A xylitol conversion of 100% was reached after 24 h, and this was accompanied by an increase in selectivities to PG (41%) and EG (30%). This performance was attributed to the existence of barium peroxide on the surface, as observed by the XPS of Ba 3d. This reduced the total basicity of the catalysts compared with the other prepared catalysts. The peroxide may have contributed to acidic sites in the dehydration reaction, which favored the formation of PG. Ni/CA had a similar reaction profile and product distribution as Ni/BA; however, after 12 h, xylitol conversion increased to 60%, and the selectivity to PG slightly increased to 35% with 30% EG. After 18 h, xylitol conversion was 78% with PG and EG selectivities at 37% and 35%, respectively. After 24 h, xylitol conversion was 96%, PG selectivity was 37.8%, and EG dropped to 30%. Over the evaluated time, there were no significant changes observed for alcohol and other products, with the selectivity of 4% for alcohol and 11% for other products. Ni/LA showed the lowest initial conversions compared to those of the other catalysts. After 6 h, it only gave 16% xylitol conversion with selectivity to glycerol at 31% and 23% to PG. No alcohols were detected, and C4 alditols were mostly produced. CO2 and methane were also produced. After 12 h, the highest glycol selectivity over Ni/LA was observed with 35% PG and 32%, respectively, to EG. At 18 and 24 h, xylitol conversion increased, but there was no change in product selectivity. This performance was surprising since Ni/LA showed the smallest Ni0 crystallite size, 10.8 nm, and the highest metal dispersion of 9.4% (Table ). This showed that the support effect plays a crucial role in xylitol hydrogenolysis. From XPS data, some carbonates were seen on the Ni/LA surface, and these may have influenced the reaction, leading to the formation of side products. Ni/MA showed better xylitol conversion (39%) after 6 h than Ni/LA, and the conversions increased to 58%, 63%, and 81% after 12, 18, and 24 h, respectively. The major difference for the Ni/MA catalyst was product distribution. EG selectivity was not stable over the period, unlike that of the other catalysts. PG showed the highest selectivity after 18 and 24 h. Other products in the liquid phase over Ni/MA comprised condensation products, esters, and carboxylic acids. In the gas phase, methane was dominant. Catalyst Structure–Activity Correlation Generally, the increasing basic strength of the prepared supports is reported as follows: La2O3 < MgO < CaO < BaO. , However, according to CO2 chemisorption data displayed in Table , it follows the sequence, Ni/CA> Ni/MA > Ni/LA > Ni/BA. This had a significant effect on the glycol yields. Figure S4 shows the PG yields of the prepared catalysts over time. Ni/BA gave higher PG yields over time compared to the other catalysts, followed by Ni/CA, then Ni/MA, and Ni/LA, showing the lowest PG yields. When looking at the EG yield, Ni/CA showed slightly higher EG yields than Ni/BA over the investigated time. There was a sudden increase in EG yield after 12 h over Ni/MA, but then it dropped after 18 h, due to the formation of other products. Ni/LA showed the lowest yields of glycols over time. Based on the data obtained, it is observed that all the catalysts showed higher PG yields compared to EG. This was attributed to the fact that when the C2–C3 bond in xylitol cleaves, two pathways can lead to the formation of PG. , PG can be formed directly from glyceraldehyde via a dehydration reaction or can be formed from glycerol. This makes PG favored compared to EG. 1,2-Butanediol was observed over Ni/BA and Ni/CA, and this evidence of C1–C2 cleavage comes from decarbonylation, producing CO and erythritol, which further reacts to form 1,2-BDO. Selectivity toward glycols and glycerol was evaluated for the catalysts that showed conversions with approximately a 5% difference. After 6 h of reaction time, Ni/MA gave 39% xylitol conversion, Ni/BA gave 41% conversion, and Ni/CA gave 44% conversion. Ni/CA showed the highest selectivity toward EG (34%), followed by PG (33%) and glycerol (24%). Over Ni/BA, PG had the highest selectivity (31%), followed by EG (26%) and glycerol (23%). Over Ni/MA, PG had 27% selectivity, EG had 20%, and glycerol was 26%. The differences in product distribution are attributed to catalyst structural differences, as discussed below. Ni/BA showed mixed phases of the oxide and spinel in its XRD pattern and hence contributed significantly to the activity. The mixed phase can facilitate the different reaction steps, such as dehydration and hydrogenolysis, leading to higher activity and selectivity toward glycols. In contrast, spinels have a rigid structure and are known to improve metal dispersion and stability, but they tend to favor a single reaction step. MA’s diffraction pattern showed that the support was a spinel, with a small amount of oxide. After impregnation, the spinel structure was retained, which led to the catalyst being dominated by medium basic sites and minor strong basic sites. The catalyst (Ni/MA) structure led to low conversion and reduced selectivity. The formation of condensation products and large amounts of CH4 was attributed to the spinel in the MA support. Metallic nickel plays a significant role in the dehydrogenation of xylitol and the hydrogenation of the resulting glycolaldehyde to form glyceraldehyde. According to TPR, the reducibility of the prepared catalysts was in similar ranges. Ni/LA had the smallest Ni0 crystallite size and the largest metal surface area, whereas there was no significant difference between the metal surface areas of Ni/BA and Ni/CA. A Lewis base is known to facilitate C–C bond cleavage, and based on the obtained data for Ni/LA, the small crystallite size and high dispersion led to excess hydrogenolysis, which favored the formation of cracked products and gases such as CH4 and CO2. Ni/BA and Ni/CA showed similar product distributions, and this was attributed to the Ni0 environment since nickel dispersion and crystallite size were in the same range. Temperature Effect The Ni/BA catalyst was chosen to study the effect of temperature on the hydrogenolysis of xylitol at a constant H2-pressure (40 bar), as it showed better activity than the other catalysts. To study the effect of temperature, a 6-h reaction time was chosen as it was expected to show the most significant changes in conversion and selectivity. Temperatures were varied between 200 and 240 °C, and the results are shown in Figure . At 200 °C, the conversion was 41% with 51% selectivity toward the combined glycols. With an increase in temperature to 220 °C, conversion reached 68%, and this was accompanied by an increase in PG selectivity from 31% at 200 °C to 43% at 220 °C. There was no significant change in EG selectivity. A conversion of 100% was obtained at 230 °C, where PG selectivity reached a maximum of 53%, and EG decreased to 26% from 28%. At 240 °C, 100% conversion was maintained; however, at this high temperature, there was a decrease in PG selectivity to 48%, and EG increased to 29%. The higher degree of C–C and C–O bond cleavage achieved at higher temperatures is attributed to good Ni0 activation. Glycerol and alcohol selectivity decreased with increasing temperature; higher temperatures favored the transformation of glycerol to PG, and alcohols degraded to form gaseous products. At 240 °C, other products were mostly composed of methane, ethane, and CO2. This data agrees with observations reported in the literature. , Effect of temperature on xylitol conversion and product selectivity over the Ni/BA catalyst. Reaction conditions: 15 wt% xylitol, 0.5 g of catalyst, 40 bar of H2 pressure. The Arrhenius equation was used to calculate the activation energy for Ni/BA, as shown by the equation. lnk=−Ea/RT+lnA 1 where k is the rate constant determined by assuming a pseudo-first-order kinetic equation, A is the pre-exponential factor, E a is the activation energy (kJ/mol), T is the temperature, and R is the gas constant (8.3145 mol–1 K–1). k can be calculated using the first-order equation lnC=−kt+lnC0 2 where C 0 and C are the initial xylitol concentration and xylitol concentration after some reaction time, respectively. The activation energy was calculated between two temperature intervals from 200 to 230 °C, holding pressure constant. Figure S5 displays the fitted line of the Arrhenius equation. The activation energy was calculated from the slope of the curve, and the y-axis provided the pre-exponential factor. The E a for xylitol hydrogenolysis over Ni/BA at 40 bar H2 pressure was calculated to be 85.02 kJ/mol. The catalyst was kinetically more favorable for xylitol activation at the evaluated parameters. Effect of H2–Pressure The effect of H2 pressure (30–50 bar) on glycol selectivity was studied at 240 °C and 6 h of reaction time. The temperature of 240 °C was chosen because 100% xylitol conversion was achieved at this temperature, and PG selectivity was comparable to that at 230 °C. The results obtained are presented in Figure . At 30 bar, high xylitol conversion was achieved (99%) with 43% PG, and 22% EG selectivity. At 50 bar, 100% conversion was retained, and PG selectivity was 49%, like that obtained at 40 bar. EG selectivity was 29%. It was noted that with an increase in H2 pressure, glycerol selectivity increased from 3% at 30 bar, 4% at 40 bar, and 6% at 50 bar. The increase in glycerol selectivity with pressure was attributed to the hydrogenation of glyceraldehyde, which is formed after C2–C3 bond cleavage in xylitol via the retro-aldol reaction, which then aided in overcoming the equilibrium limitation associated with dehydrogenation reactions due to product accumulation. This was corroborated by the disappearance of the glyceraldehyde intermediate at 50 bar compared to 30 bar, as found by GC-MS. Methane and CO2 formation was suppressed at 50 bar, and this was attributed to the hydrogenation of intermediates to form stable alcohols. Effect of the pressure on the cytopoly­(xylitol) conversion and product selectivity over the Ni/BA catalyst. Reaction conditions: 15 wt% xylitol, 0.5 g catalyst, Temp = 240 °C. Catalyst Loading Effect The influence of mass diffusion on the conversion of xylitol was evaluated using different masses of Ni/BA at 230 °C and 40 bar H2 pressure for 6 h at fixed xylitol concentrations. Figure shows a linear relationship between conversion and catalyst mass; 0.1 g of catalyst showed low xylitol conversion at 44% with good selectivity toward the glycols. When the mass was increased to 0.2, 0.3, and 0.5 g, xylitol conversions increased. Increasing the catalyst mass increases the number of active sites for xylitol hydrogenolysis, leading to higher conversions. At low catalyst mass, the catalyst may be easily saturated, which leads to low xylitol conversions. There was no significant change in glycol selectivity when using catalyst masses of 0.1, 0.2, and 0.3 g. High glycol selectivity was obtained when using 0.5 g of catalyst. This was attributed to increased active sites for hydrogenolysis and hydrogen dissociation, which transform C4 and C3 aldose and ketose products into PG, EG, and glycerol. A catalyst loading of 0.5 g also gave the lowest amounts of alcohols and other products, showing limited cracking or glycol degradation. Effect of catalyst loading on xylitol conversion and product selectivity over the Ni/BA catalyst. Reaction conditions: 15 wt% xylitol, 40 bar H2 pressure, Temp = 240 °C. Insights into Reaction Pathways The product distribution observed over Ni/BA in the hydrogenolysis of xylitol showed that a series of parallel reactions is taking place, with a plausible reaction pathway proposed in Scheme . Hydrogenolysis of xylitol occurs over the catalyst surface, where the catalyst facilitates C–C and C–O bond cleavage. This proceeds by multistep reactions involving dehydrogenation (DH), hydrodeoxygenation (HDO), hydrogenation, and partial DH reactions. According to the literature, carboxylic acids are formed by partial dehydrogenation of xylitol. , The observation of butanoic acid and acetic acid supports partial DH. Acetic acid is formed from PG through C–C bond cleavage. Xylitol DH can lead to aldose or a ketose intermediates; aldose leads to C1–C2 cleavage, resulting in decarbonylation and the formation of CO and erythritol, which further reacts to produce 1,2 BDO and butanoic acid. The selectivity to aldose intermediates is low at high H2 pressure when using a Ni catalyst. This is shown by low amounts of C4 products obtained and low selectivity to CO. High hydrogen pressure favors the formation of the ketose intermediate. This was seen by high selectivity to C2 and C3 products, as shown by high PG, EG, and glycerol selectivity and yields. Proposed Reaction Pathway for Xylitol Hydrogenolysis over Ni/BA Recyclability Studies of Ni/BA Recyclability studies were conducted over Ni/BA as the best-performing catalyst, and the obtained results are displayed in Figure . As observed, Ni/BA exhibited a loss of activity over the four repeated cycles. After each run, the catalyst was washed with methanol and dried at 110 °C under vacuum to remove physisorbed organic species. It was not possible to perform catalyst regeneration by washing with acid, as this leads to metal (Ni) leaching and catalyst mass loss. Xylitol conversion decreased from 100% to 93% after the first cycle, and this was accompanied by a decrease in PG selectivity from 53% to 49%. There was a further decrease in xylitol conversion in the subsequent cycles, with a drop to 72% after the third cycle. There was a shift in product distribution with a significant decrease in glycol selectivity and an increase in other products, which were composed mainly of methane. Selectivity toward alcohols remained unchanged. The C4 products and carboxylic acids also increased. A decrease in activity was attributed to carbonaceous deposits over the catalyst surface, which resulted in the blocking of active sites. Reduced active sites resulted in low xylitol and hydrogen adsorptions over the catalyst surface, which was required to facilitate the reaction. Reduced active sites promoted selectivity toward carboxylic acids due to the limited access to hydrogen. Carbon deposits were identified using TGA-DSC, as shown in Figure . The appearance of a sharp peak around 300 °C in the spent catalyst was indicative, as this peak was not present in the fresh catalyst. When looking at the spent catalyst, there was a strong exothermic peak at around 300 °C. This was attributed to the oxidation of organic species or the burning of coke present on the catalyst surface. This aligns with the sharp mass loss step. Heat flow then went back to the baseline after 350 °C, showing the completion of the exothermic reaction. At high temperatures around 800 °C, the heat flow steadily returned to endothermic, as evidenced by the absence of sharp peaks, indicating that no chemical reaction was taking place. The fresh catalyst showed a small endothermic effect around 150 °C with no oxidation present. Recyclability studies over Ni/BA using 0.5 g of catalyst, 15 wt% xylitol at 230 °C, 40 bar H2-pressure over 6 h. TGA–DSC curves under air flow showing weight loss and heat flow for fresh Ni/BA and spent Ni/BA after 4 cycles. Conclusions A series of bifunctional catalysts was prepared to study the influence of modifying the support by the incorporation of alkaline metals in alumina. XRD revealed that some of the supports were spinels, while others were amorphous, and all were mesoporous, according to the N2 physisorption isotherms obtained. Ni showed a high degree of reducibility in all the prepared catalysts with low dispersion, which may be attributed to Ni0 sintering due to high reduction temperatures and the method of catalyst preparation. XPS revealed that even though the alkaline metals were in the same group, they interacted differently with alumina, as some formed peroxides with oxygen, and others were simple oxides. All the prepared catalysts were active in the hydrogenolysis of xylitol, and incorporating Mg2+, Ca2+, Ba2+, and La3+ had a significant promotional effect on catalyst performance. Ni/BA and Ni/CA gave better conversions, achieving 100% and 96% after 24 h, and high yields toward the desired glycols. C2–C3 hydrogenolysis was found to be more favored than C1–C2 cleavage. This was confirmed by high C3 and C2 product yields, while C4 product yields were low due to the good interaction between oxygen and the alkaline metal, which increases the number of moderate basic sites. Reaction kinetics showed that first-order reactions were followed. Xylitol activation was found to have a relatively low activation energy when using Ni/BA. Experimental Section Support and Catalyst Preparation Supports were prepared via the coprecipitation method with a target mole ratio of 3:10 metal oxide (MO) to Al2O3. In a typical synthesis, the required amount of alkaline earth metal nitrate salts and Al­(NO3)3·9 H2O (ACS reagent, ≥98%) were added to deionized water in a beaker and dissolved at 60 °C under vigorous stirring. Ammonium hydroxide solution was added dropwise at constant stirring until precipitation was reached at pH 10. Then, the solution was heated to 80 °C and kept there for 5 h. The precipitate was filtered off and washed with hot water until a neutral pH was reached. The product was then oven-dried overnight at 110 °C. The obtained solid was ground with a mortar and pestle to form a powder, which was calcined in a tube furnace under a flow of air at 700 °C, obtained at a heating rate of 10 °C/min, for 5 h. Then, the contents in the furnace were cooled naturally by the flow of air. The prepared supports are denoted as follows: MgO-Al2O3 (MA), CaO-Al2O3 (CA), BaO-Al2O3 (BA), and La2O3–Al2O3 (LA). NiO was prepared using a similar method using sodium hydroxide as a precipitating agent. The supported Ni catalysts with a nominal loading of 15 wt% Ni were prepared by impregnating the prepared supports with an aqueous solution of Ni­(NO3)2·6H2O (Sigma-Aldrich, 99.9%). The excess water was evaporated using a rotary evaporator over 8 h in a temperature range of 50–85 °C at constant pressure. The obtained powder was then calcined under air at 350 °C, obtained by heating at 10 °C/min, for 4 h to obtain the final catalyst. The catalysts are denoted as follows: Ni/MA, Ni/CA, Ni/BA, and Ni/LA. Catalyst Characterization Brunauer–Emmett–Teller (BET) surface area and pore volume measurements were carried out using a Micrometrics Tristar II Surface Area and Porosity Analyzer. Before analysis, the samples were weighed and degassed overnight at 250 °C under a constant flow of nitrogen gas. Inductively coupled plasma (ICP-OES) analysis was performed using a PerkinElmer Optical Emission Spectrometer Optima 5300 DV. Before analysis, samples were acid-digested using HNO3 and then diluted accordingly. The crystallite phases were determined by XRD conducted on a Bruker D8 Advance diffractometer with Cu (Kα, λ = 1.5406 Å) as the radiation source. Data evaluation was performed by using the EVA software from Bruker. XPS analysis was performed on an ESCALAB 250 X-ray photoelectron spectrometer (Thermo-Fisher) equipped with Al Kα1,2 monochromatized radiation at a 1486.6 eV X-ray source. Charge neutralization was achieved using a combination of low-energy electrons and argon ions, resulting in a reproducible C­(1s) binding energy of 284.8 eV for all prepared samples. Hydrogen TPR analyses were carried out on a Micromeritics Autochem 2920 chemisorption analyzer. Before reduction, approximately 60 mg of the sample was placed in a U-shaped quartz tube. The sample was pretreated by heating at 5 °C min–1 under helium flow (50 mL min–1) to 120 °C and kept at this temperature for 10 min. The sample was then cooled to room temperature and subsequently heated again at 10 °C min–1 under 100% H2 flow (50 mL min–1) to 1000 °C and kept at this temperature for 10 min. The water formed during reduction was trapped using a dry ice/isopropanol bath. The amount of hydrogen consumed during reduction was measured with the TCD. CO2–temperature-programmed desorption-MS (TPD-MS) experiments were carried out on a Micromeritics Autochem 2920 chemisorption analyzer coupled to a Cirrus Mass Spectrometer. Approximately 60 mg of the sample was placed in a quartz U-tube containing quartz wool, fitted with a thermocouple for continuous temperature measurements. Before the TPD experiments, the samples were degassed at 500 °C for 60 min in a flow of helium (50 mL min–1) at 10 °C min–1. Then, the sample was cooled to 80 °C. The activation of the samples was followed by pulse chemisorption, where 30 injections of CO2 gas were dosed onto the catalyst using an inert gas, helium, flowing through a 1 cm3 loop. The last part of the analysis involved TPD, which was carried out with a linear heating rate (10 °C min–1) in a flow of helium (50 mL min–1) from 100 to 900 °C for 60 min. Particle morphologies of the prepared catalysts were investigated with transmission electron microscopy (TEM) images using a Jeol JEM-1010 Electron Microscope. The images were captured and analyzed using iTEM software. Before the analysis, the catalysts were sonicated in ethanol, after which they were placed on a copper grid for analysis. Nickel dispersion on the catalyst supports was analyzed by hydrogen chemisorption using a Micromeritics ASAP 2020 Chemisorption analyzer. Catalysts were degassed overnight under a vacuum at 250 °C. Before the measurements, catalyst samples were reduced in flowing hydrogen (50 mL min–1) at 400 °C, and hydrogen was passed through the sample at a temperature of 200 °C with increasing pressure until saturation of the sample was attained. Upon completion of the analysis, the sample was flushed with helium to terminate the analysis. Catalytic Testing Before catalytic testing, the prepared catalysts were reduced with H2 (99.999%) flowing at 50 mL min–1 (controlled by a Bronkhorst mass flow controller) at 400 °C for 4 h at atmospheric pressure. After reduction, the reactor was cooled to room temperature under N2 (99.999%) gas flowing at 20 mL min–1, and the catalyst was passivated with 1% v/v O2/N2 at room temperature. Hydrogenolysis of xylitol was carried out in a Parr autoclave (100 mL) at a stirring speed of 1500 rpm. In a typical run, 40 mL of 15 wt% xylitol solution and 0.5 g of catalyst were introduced into the autoclave. After purging with N2, the reactor was heated to the reaction temperature (200–240 °C). Then it was pressurized to 40 bar with H2. After the reaction, the reactor was cooled to below 10 °C. Gas samples were collected first for analysis, and the reactor was then depressurized to atmospheric pressure. The catalyst was separated by centrifuging the sample mixture and filtering through PVDF membranes (0.45 μm). Gaseous products (CO2 and CO) were analyzed using a PerkinElmer Clarus 400 Gas Chromatograph fitted with a 30 m × 530 μm Supelco Carboxen 106 PLOT column and a thermal conductivity detector (TCD). Liquid and other gaseous products, such as CH4, were analyzed using a Shimadzu GC-2025, fitted with a 50 m × 200 μm PONA capillary column and a flame ionization detector (FID). The injection temperature was 240 °C for an injection volume of 0.2 μL with a split ratio of 150. Conversion, selectivity, and carbon balance were determined as follows: Conversion=Cinitialmolxylitol−CfinalmolxylitolCinitialmol−xylitol×100 Selectivity=molesofproducttotalmolesofproducts×100 Carbonbalance=CfinalmoleproductCconvertedmolxylitol×100 Supplementary Material