Catalytic hydrogenation of stearic acid to 1-octadecanol using supported bimetallic Pd–Sn(3.0)/γ–Al2O3 catalyst

Received: 4th January 2022 Revised: 20th February 2022 Accepted: 24th February 2022 Online: 28th February 2022


Introduction
The depletion of fossil fuel pushed numerous scientists to have great attention to searching for new energy resources and strategies to meet the increasing energy demand. The production of biodiesel has been intensively studied in recent years, in part due to political decisions to increase the use of biofuels, especially biodiesel palm oil-based. For instance, the Indonesian government has set a target energy level mixed with B50 to fossil diesel by 2025 [1]. On the other hand, diesel-like hydrocarbons, consisting of C16 to C18 carbon atoms, provide good fuel properties, such as viscosity, cloud point, and boiling point [2]. Long-chain hydrocarbons can be produced from fatty acids and their derivatives via hydrotreatment or catalytic hydrodeoxygenation (hydrogenation-dehydrationdecarbonylation) [3]. The hydrogenation of fatty acids (e.g., stearic acid) to fatty alcohol using both heterogeneous and homogeneous catalysts is the crucial step in the transformation of biobased resources since fatty alcohol is widely used as the component of cosmetics, food ingredients, surfactant, plasticizer, lubricant, and intermediate of biofuel synthesis [4]. The use of homogeneous catalysts, such as Ru-Triphos complex [5] or iron-base PNP-pincer [6], showed high selectivity towards alcohols from the hydrogenation of carboxylic acids. However, homogeneous catalysts demonstrated much more selective than heterogeneous ones, with several drawbacks such as using expensive organic ligands, lack of reusability, and challenges associated with removing residual heavy metals in the isolated products. However, due to the high cost of the metal ligands used in these catalyst systems and problems associated with recycling the homogeneous catalysts, heterogeneous catalysts have proven to be more attractive choices for industrial applications [7].
Several attempts on the hydrogenation of stearic acid to 1-octadecanol using heterogeneous catalysts have

)
Check for updates been reported previously [8]. The hydrogenation of fatty acids over a copper-chromite catalyst is commercially used to hydrogenate fatty acid esters to alcohols under severe conditions (2-50 MPa, 473-673 K) [9,10,11]. In addition, the presence of toxic chromium in the copperchromite catalyst poses an environmental and health hazard. Therefore, the development of catalysts without using Cr and under milder reaction conditions has been attempted. Bimetallic platinum group metal (PGM) based catalysts such as Pt-Re/TiO2 (2.0 MPa, 403 K) reported 61-90% selectivity to C10-C18 fatty alcohols at 79-83% conversion [12]. The presence of co-promotor ReOx in Pd-ReOx/SiO2, Rh-ReOx/SiO2, and Ir-ReOx/SiO2 catalysts (8.0 MPa, 413 K) enhanced the performance of Pd, Rh, and Ir catalysts to have 94-98% selectivity of 1octadecanol at 100% conversion [13]. However, noble metal-based catalysts and low substrate loading are not economical and have less viability in upgrading the biomass-derived platform industry. Therefore, alternative economic and eco-friendly heterogeneous catalysts that would ensure the preferred hydrogenation of the carboxylic acids (fatty acid) to fatty alcohol are highly desired.
The electropositive metals such as tin (Sn), indium (In) of iron (Fe) have been widely used as co-promotor for Ru, Pd, or Ni-based bimetallic catalysts. Toba et al. reported that 2wt% Ru-4.7wt% Sn/Al2O3 showed high selectivity toward alcohols (89.4%) at 97.3% adipic acid conversion at 513 K and 6.5 MPa of H2 [14]. Bimetallic Ni-Sn alloy supported on TiO2 (Ni-Sn(1.5)/TiO2) catalyst showed a high yield of lauryl alcohol at (97%) at >99% conversion of lauric acid at 433 K, 3.0 MPa H2 for 20 h [15]. Damayanti et al. reported using bimetallic Pd-Fe/TiO2 catalyst for the hydrogenation of levulinic acid (LA) to γ-valerolactone (GVL) at 443 K, 3.0 MPa H2 for 7 h. The obtained yield of GVL over Ru-Fe/TiO2 was 52.4%, much higher than that of Pd/TiO2 (22.7%) [16]. Most recently, supported bimetallic Pd-Sn(x)/C catalysts demonstrated a high yield of 1-octadecanol (73.2%) at 100% conversion of stearic acid. The high selectivity of alcohols over Pd-Sn(1.5)/C catalyst can be attributed to the formation of bimetallic Pd-Sn alloy phases (e.g., Pd3Sn and Pd3Sn2) as obviously depicted by XRD analysis [17]. Scheme 1. Conceived reaction routes for the catalytic transformation of stearic acid using heterogeneous catalysts In this paper, an extended investigation on the gamma-alumina supported bimetallic palladium-tin catalysts (denoted as Pd-Sn(3.0)/γ−Al2O3; Pd = 5%wt and Pd/Sn molar ratio is 3.0) for the hydrogenation of stearic acid to 1-octadecanol under mild reaction conditions has been investigated. The current results showed that Pd-Sn(3.0)/γ−Al2O3 catalyst showed the highest conversion of stearic acid (99.1%) and 1-octadecanol yield (43.2%), which were higher than that of other supported Pd-Sn catalysts under the applied reaction conditions.

Catalyst preparation
A typical procedure of the synthesis of supported bimetallic Pd−Sn(3.0)/γ−Al2O3 (Pd = 5%b/b and Pd/Sn feeding molar ratio of 3.0) catalyst is described as follows [15,17]: Pd(CH3COO)2 (0.4613 mmol) was dissolved in deionized water (denoted as solution A), and SnCl2·2H2O (0.1537 mmol) was dissolved in ethanol/ethylene glycol (20:10 v/v mL) (denoted as solution B) at room temperature. Solutions A and B and 1.0 g of support (γ− Al2O3) were mixed at room temperature; the temperature was subsequently raised to 323 K, and the mixture was gently stirred for 12 h. The mixture was adjusted to pH 12 by dropwise an aqueous solution of NaOH (3.1 M or 6.0 M) and placed into a sealed-Teflon autoclave for the hydrothermal reaction at 423 K for 24 h. The resulting black precipitate was filtered, washed with distilled water, and then dried under vacuum overnight. Prior to the catalytic reaction, the obtained black powder was reduced with hydrogen (H2) gas at 673 K for 3 h [18]. Bimetallic Pd−Sn supported on Y-Zeolite, HZSM-5 Si/Al = 8, Nb2O5 dan TiO2 catalysts were also synthesized using a similar procedure.

Characterizations
The prepared catalysts were characterized by powder X-ray diffraction on a RIGAKU MINIFLEX 600 instrument using monochromatic CuKα radiation (λ=0.15418 nm). It was operated at 40 kV and 20 mA with a step width of 0.02° and a scan speed of 5° min -1 . According to Scherrer's equation, the mean crystallite size of Ni was calculated from the full width at half maximum (FWHM) of the Pd (111)  Heptadecane specific surface area employing the BET equation. The total pore volume was estimated to be the liquid volume of nitrogen at a relative pressure of about 0.995. The Barrett-Joyner-Halenda (BJH) approach was used to calculate desorption data's total pore volume and size distribution [19].
The temperature-programmed desorption of ammonia (NH3-TPD) was carried out on a Belsorp Max (BEL Japan). The samples were degassed at an elevated temperature of 373−473 K for 2 h to remove physisorbed gases prior to the measurement. The temperature was then kept at 473 K for 2 h while flushed with helium gas. NH3 gas (80% balanced NH3 and 20% He) was introduced at 373 K for 30 minutes, then evacuated by helium gas to remove the physisorbed for 30 minutes. Finally, temperature-programmed desorption was carried out at 273-1073 K temperatures, and TCD monitored the desorbed NH3.
Attenuated total reflection-Fourier transformed infrared (ATR-FTIR) analysis was performed on a Diamond Bruker spectrometer with a resolution of 2 cm -1 and a scanning number of 36. A 20 mg sample in powder form was degassed under vacuum (10 −3 Pa) for 1 h at 423 K in the TAIATSU techno with glass-fitted inside. Then, it was cooled to room temperature, and the initial background spectrum was recorded. After the sample was exposed to pyridine vapor under vacuum for 60 minutes and overnight, followed by removal of the excessive pyridine at 323 K for 0.5 h, then measured by ATR-FTIR.

Catalytic reactions
A typical procedure for hydrogenation of stearic acid was described as follows: catalyst (0.05 g), stearic acid (0.2844 g; 1.0 mmol), 2-propanol:H2O (5 mL; 4.0:1.0 v/v) as solvent were placed into a glass reaction tube, which fitted inside a stainless steel reactor. After H2 was introduced into the reactor with an initial H2 pressure of 2.0 MPa at room temperature, the reactor's temperature was increased to 513 K using an electric furnace, 800 rpm. After 13 h, at room temperature, the internal standard of dodecane was added, and the conversion of stearic acid and the yield of 1-octadecanol were determined by GC analysis. The Pd−Sn/γ−Al2O3 catalyst was easily separated using simple centrifugation or filtration.

Product Analysis
GC analysis of the reactant (stearic acid) and products (1-octadecanol, ester, and heptadecane) was performed on a Perkin Elmer AutoSystem XL equipped with a flame ionization detector and Thermo Scientific (0.25 mm × 15 m × 0.25 μm) capillary column. It was operated under the following conditions: injector and detector temperatures (523 K); airflow (450 mL/min); H2 flow (45 mL/min); N2 flow (14 mL/min); and a split ratio of 50:1. The temperature column has been set gradually into two steps (first: 373-493 K (ramping of 20 K/min) and second: 493-573 K (ramping of 18 K/min). Gas chromatography-mass spectrometry (GC-MS) was performed on a Shimadzu GC-17 equipped with a thermal conductivity detector and an RT-βDEXsm capillary column. The products were confirmed by comparing their GC retention time and mass spectra with those of authentic samples.
The conversion, yield, and selectivity of the products were calculated according to the following equations: Conversion: F0 is the introduced mol reactant (stearic acid), Ft is the remaining mol reactant, and ΔF is the consumed mol reactant (introduced mol reactant−remained mol reactant), which are all obtained from GC analysis using a standard internal technique.
The hysteresis loop of adsorption/desorption of the synthesized bimetallic Pd−Sn(3.0)/γ-Al2O3 sample is very similar to that of IV type, which indicates the strong interaction between the molecule adsorbate and catalyst surface. The formation of the hysteresis loop also indicated the condensation of molecule adsorbate during the desorption of N2 gas [20] (Figure 2(a)). The plot of the volume of adsorbed-N2 versus pore distribution using the Horvath-Kawazoe (HK) method was performed to determine the pore size distribution of the synthesized catalysts, as shown in Figure 2(b). The pore size distribution was ≥1.16 nm after reduction with H2 at 400°C. However, there is no clear evidence for the change in the pore size distribution towards small pore sizes or big pore sizes after introducing the Pd−Sn species or thermal activation using N2 or H2 at 400°C.  To validate the pore structure (micropore and mesopore area) and the possible adsorption of the catalyst sample, the t-plot and α-plot techniques were performed. The results are shown in Figure 3. The frontline of t-plot profiles showed that the sample has mainly micropore structure (≤ 1.0 nm) (Figure 3(a)), whereas the frontline of α-plot profiles showed both the micropore and the mesopore structures (≥ 2.0 nm) (Figure 3(b)). The pore size at ≥ 2.0 nm was observed when the volume of N2 gas increased up to 200 cm 3 /g. Therefore, it can be concluded that the synthesized Pd− Sn(3.0)/γ-Al2O3 catalyst has both micro and mesopore structures [21].  Table 1 summarized the porosity properties (specific surface area (SBET), micro and mesopore surface area, pore-volume, and pore size distributions (entry 1). Pd-Sn(3.0)/γ-Al2O3 catalyst has similarity with Pd-Sn(3.0)/C, which has both microstructures on the surface. The differences are Pd-Sn(3.0)/C has a higher surface area than Pd-Sn(3.0)/γ-Al2O3 with pore size distribution and pore volume smaller than Pd-Sn(3.0)/γ-Al2O3 (entry 2) [17]. The total acid site density was derived from the amounts of desorbed ammonia, which were formally divided into three temperature regions to denote three types of acid sites: (1) weak acid sites, ranging from 373 to 573 K, (2) moderate acid strength, ranging from 573 to 823 K, and (3) high acid strength, ranging from 823 to 1023 K [22,23,24,25]. The results of temperatureprogrammed desorption of ammonia (NH3-TPD) and ATR-FTIR of pyridine adsorption are shown in Figure 4 and Table 2. Pd−Sn(3.0)/γ-Al2O3 catalyst has weak and strong acid sites with amount acid sites 11 µmol/g and 380 µmol/g, respectively. ATR-FTIR analysis of adsorbed pyridine also confirmed that the Pd−Sn(3.0)/γ-Al2O3 catalyst has both Lewis and Brönsted acid sites, suggesting that the effect of total acidity might play an essential role during the selective hydrogenation of stearic acid to 1-octadecanol (stearyl alcohol) under the current operating conditions.   Figure 5 shows the XRD patterns of commercial γ-Al2O3 support, as prepared and pre-reduced bimetallic Pd−Sn(3.0)/γ-Al2O3 catalyst. The commercial γ-Al2O3 support exhibited a typical diffraction peak of crystalline of γ-Al2O3 support (Figure 5(a)). In the case of Pd− Sn(3.0)/γ-Al2O3 catalyst, the typical diffraction peaks at 2θ = 14°; 32.9°; 34.2°; dan 65.9° were clearly observed, which can be attributed to the tin oxide (SnO) (JCPDS#04-673), Sn (101) (JCPDS#18-1380) and SnO2(301) (JCPDS#41-1445) phases, respectively ( Figure  5(a)(b)) [7,26,27]. A high intensity of the diffraction peak at 2 = 40.2° of Pd−Sn(3.0)/γ-Al2O3 sample after reduction with H2 at 673 K that can be suggested due to the modified surface structure of Pd(111) in the presence of Sn promoter either to form surface bimetallic or alloy Pd−Sn (Pd2Sn species; JCPDS#007-1070) ( Figure 5(c)) [28,29,30].

Screening of catalyst
The catalytic reaction of stearic acid using various palladium-based catalysts was performed in 2propanol/H2O (4.0:1.0 v/v) solvent, 513 K, 2.0 MPa H2 for 7 h, and the results are summarized in Table 3. The catalytic conversion of stearic acid using various supported bimetallic Pd−Sn catalysts was investigated, and the results were also shown in Table 3. The conversion of stearic acid was varied at 47-92%, the maximum yield of 1-octadecanol was only 6.5% at a conversion of 91.7%, obtained over Pd−Sn(3.0)/HZM-5 catalyst (entries 1-4). The currently obtained yields of 1octadecanol using Pd−Sn(3.0) supported on various metal oxides and zeolites were much lower than that of supported Pd−Sn(1.5) catalysts under similar reaction conditions as recently reported by Rodiansono et al. [17]. By using Pd−Sn(3.0)/γ-Al2O3 catalyst, a remarkably high yield of 1-octadecanol (22.7%) at 94.7% conversion of stearic acid was achieved (entry 5).
In order to understand the reaction profiles, the hydrogenation of stearic acid in the presence of a Pd− Sn(3.0)/γ-Al2O3 catalyst at different reaction times was carried out. At an earlier reaction time of 1 h, the conversion of stearic acid was only 12.9% without forming a hydrogenated product of 1-octadecanol (entry 6). The yield of 1-octadecanol significantly increased to 12.5% (at 84.4% conversion of stearic acid) when the reaction time of 5 h was applied. Further prolonged reaction time to 13 h, the conversion of stearic acid was nearly 100%, and a maximum yield of 1-octadecanol (43.2%) was obtained (entry 8). To confirm the importance of bimetallic Pd-Sn instead of monometallic Pd catalyst, the catalytic reaction over Pd/γ-Al2O3 gave 81.3% conversion of stearic acid. The products were distributed to 1-octadecanol (2.1%), ester (isopropyl stearate) (26.5%), and others (52.7%) after 13 h (entry 9). These results suggested that the presence of Sn in Pd− Sn(3.0)/γ-Al2O3 might enhance the selectivity product of 1-octadecanol by inhibiting the decarboxylation reaction compared to the monometallic Pd/γ-Al2O3 catalyst [31,32]. The conversion of stearic acid using Ni-FeOx and Ru3Sn7/SiO2 catalysts were reported as 75-100%. The 20% and 99% yields were obtained using Ni-FeOx and Ru3Sn7/SiO2 catalysts, respectively (entry [10][11]. Although both catalysts showed good performance, these catalysts required harsher reaction conditions than the Pd-Sn/γ-Al2O3 catalyst.   Reaction conditions: catalyst (0.2 g); solvent (dodecane 80 mL); temperature 513 K; initial H2 pressure (4.0 MPa) [7]. d Conversion was determined by GC using a standard internal technique. d Yields were determined by GC using GC area ratio according to GC-MS data. d Others were included hydrocarbon and unidentified products based on the GC-MS analysis data [17].

Effect of reaction temperature
The effect of reaction temperature on the conversion and product distribution in the catalytic conversion of stearic acid over bimetallic Pd−Sn(3.0)/γ-Al2O3 catalyst was investigated. As shown in Table 4, at the reaction temperature of 433 K, the conversion of stearic acid was 22.8%, and the products were distributed to 1octadecanol (1.1%), ester (19.1%), and other products (2.6%) (entry 1). The increase in reaction temperature to 473 K increased stearic acid conversion (49.6%). However, the yield of 1-octadecanol was remained low (2.0%), whereas the formation of ester and other products increased significantly (entry 2). When the reaction temperature was increased to 513 K, high conversion of stearic acid (94.7%) with remained low 1octadecanol yield (22.7%) was obtained (entry 3).
In these reaction conditions, the yield of ester (mainly contains isopropyl stearate) and others were still much higher than that of Pd−Sn(1.5)/γ-Al2O3 catalysts (entry 4). The hydrogenation of fatty acid using Pd-based catalysts required a co-promotor to enhance alcohol selectivity and inhibit the decarboxylation reaction, producing alkanes as the main products [13,33]. This result suggested that the further reaction of the formed alcohol to another side product occurred during the hydrogenation of stearic acid, as indicated by the number of others. Reaction conditions: catalyst (0.05 g); stearic acid (0.2844 g; 1.0 mmol); solvent (2-propanol/H2O (5.0 mL; 4.0:1.0 v/v); 513 K; initial H2 pressure (3.0 MPa); reaction time (7 h). a Conversion was determined by GC using a standard internal technique. b Yields were determined by GC using GC area ratio according to GC-MS data. c Others were included hydrocarbon and unidentified products based on the GC-MS analysis data. d Data was taken from the cited reference of Rodiansono et al. [17]. ATR-IR analysis of the recovered catalyst was carried out, and the results are shown in Figure 6. Two absorption peaks at wavenumber (υ) 1067 cm -1 and 3297 cm -1 can be assigned as C−O and O−H stretching, respectively. The presence of a sharp peak at 2970 cm -1 , which can be assigned as −CH2, was also clearly observed. A small peak at 1735 cm -1 was observed in Figure 6b, which can be attributed to the C=O functional group of the remaining stearic acid reactant. Wavenumber (cm 1 )

Conclusion
We described the selective hydrogenation of stearic acid to corresponding alcohol using Pd−Sn(3.0)/γ-Al2O3 catalyst under mild reaction conditions. The highest yield of stearyl alcohol (1-octadecanol) (up to 43.2%) at 99.1% conversion of stearic acid at temperature 240°C, initial H2 pressure of 2.0 MPa, a reaction time of 13 h, and in 2-propanol/water (4.0:1.0 v/v) solvent. Under the current reaction conditions, the main product obtained over Pd−Sn(3.0)/γ-Al2O3 catalyst was ester stearate (isopropyl stearate) and others. The high yield of ester and other products may be due to the high acidity of catalyst as indicated by NH3-TPD and ATR-FTIR pyridine adsorption analysis.