selective aqueous-phase hydrogenation of levulinic acid to pentanediol over bimetallic catalysts: the influence of the promoter
Résumé
Nowadays, a large range of acids are produced from biomass either through chemical or biochemical transformations. Levulinic acid (LevA) can be cost-effectively produced from lignocellulosic biomass via a simple acid-catalyzed hydrolysis process (Biofine technology [1]). The hydrogenation of LevA over supported metallic catalysts leads to the formation of g-valerolactone (GVL), 1,4-pentanediol (PenDO) and methytetrahydrofurane (MTHF). The selectivity depends on the nature of the metal and on the reaction conditions.
While GVL was predominantly formed over monometallic catalysts, PenDO and MTHF were mainly produced over bimetallic Re-Ru catalysts [2]. Since partial leaching of Re was observed during the aqueous-phase hydrogenation of succinic acid [3], different other promotors (Mo, Sn, W) were evaluated in the hydrogenation of LevA.
Materials and Methods
Bimetallic catalysts were prepared by successive impregnations of an active carbon (L3S CECA) with aqueous solutions of the metallic salts and reduction under H2. Hydrogenation reactions were conducted in a 300 mL Hastelloy Parr 4560 autoclave equipped with an electrically heated jacket, a turbine stirrer with magnetic driver, and a liquid sample line. In a typical reaction, the reactor was loaded with 7.5 g acid, 142 g water and 1 g catalyst. After purging with Ar, the reaction medium was heated to the desired temperature (typically 140°C) and H2 pressure (150 bar) was introduced to initiate the reaction. The liquid samples periodically withdrawn from the reactor were analyzed using both HPLC chromatography with UV and RID detections (ICSep Coregel 107H column, 0.005 N H2SO4 as mobile phase at a flow rate of 0.5 mL.min-1) and gas chromatography (FFAP column, using He as carrier gas). Acids, lactones, diols, cyclic ethers and by-products (propionic, butyric and acetic acids, n-butanol, and n-propanol) could be quantified. Total Organic Carbon (TOC) was also measured using a Shimadzu TOC-VCHS analyzer.
Results and Discussion
Regardless the nature of the promotor, complete hydrogenation of LevA to GVL was observed within less than one hour. The reaction rate of the subsequent hydrogenation of GVL was highly dependent on the Ru/promoter ratio as well as on the nature of the promoter.
In the presence of Sn-Ru/C, the hydrogenation of GVL was very slow; after 30 h only a 16% PenDO yield was obtained. Similarly, modest reaction rate was achieved in the presence of W-Ru/C catalysts. Conversely, interesting results were observed in the presence of Mo-Ru/C catalysts (Figure 1). Using a Mo/Ru molar ratio in the range from 0.26 up to 1.8, it was shown that the higher the Mo/Ru ratio, the lower the reaction rate. However, the selectivity to PenDO was not affected and was as high as 75-80%. Furthermore, no Mo leaching was detected during the reaction.
Figure 1: Evolution of the concentration of the products during the hydrogenation of LevA (not shown) over 0.5%Mo-2%Ru/C (dotted line) and 3.2%Re-2.5%Ru/C (full line)
Since formic acid is co-produced in equimolar amount as LevA during the industrial production of LevA, its presence in the feed was studied. Over both catalysts (Re and Mo promoted), a dramatic decrease in the reaction rate (together with hydrogenation of formic acid) was observed. However, the final selectivity to PenDO was unchanged. The higher stability of Mo-based catalysts compared to Re-based catalysts was confirmed.