Catalytic Conversion Of Biomass-derived Levulinic Acid To Valuable Chemicals

To address the crirical issues of the looming energy crisis, tremendous global efforts have been devoted to the utilization of biomass resources. Lignocellulosic biomass is currently the major component of renewable resources, which is readily available and does not compete with the food supply. An effective strategy for utilization of biomass is to first partially remove oxygen to produce reactive intermediates, denoted as platform molecules, followed by conversion of these molecules into desired products. The production of equimolar quantities of levulinic acid (LA) and formic acids (FA) can be achieved, in good yields, from cellulose through hydrolysis with dilute sulfuric acid. LA has been identified as one of the most attractive12platform molecules by US Department of Energy in2004. Among the top12value-added chemicals available from biomass, only LA can be produced using exclusively acid-catalyzed chemical processing, e.g., dehydration.Among the many useful chemicals can be produced from LA, γ-valerolactone (GVL) has been identified as one of the most important renewable intermediate. GVL may be used as a solvent, food additive and as a biofuel, for instance as a substitute of ethanol in gasoline-ethanol blends. In addition GVL can be converted to1,4-pentanediol (1,4-PDO) or2-methyltetrahydrofuran (2-MTHF) which are important raw materials. Despite the considerable efforts dedicated to the syntheis of GVL, some technical barriers remained to be resolved. For example, during the production of GVL from direct cellulose hydrolysis, it is not yet possible to recover and resue the residual sulfuric acid after cellulose hydrolysis. Morevoer, it is highly challenging to use the co-produced formic acid during cellulose hydrolysis as sole in situ hydrogen source for reduction of LA. In addition, there is still a great need to new readily available, and nobel-metal-free catalyst systems that can allow flexible or tunable transformation of bio-derived LA into1,4-PDO or2-MTHF via intermediate formation of GVL. Aiming to address above key issues in the field of LA utilizaiton, the following work was carried out in the present dissertation:1. Hydrogenation of LA to GVL under mild conditionsA series of supported iridium catalysts prepared by conventional impregnation have been applied for the hydrogenation of LA to GVL. Among the various supported catalysts, CNT supported iridium (Ir/CNT) catalyst has the best LA hydrogenation activity. Thus, at50℃and2MPa hydrogen pressure, an excellent GVL yield of99%can be obtained after1hour reaction. Transmission electron microscopy (TEM) analysis of the Ir/CNT catalyst reveals that the particles corresponded to metallic Ir0with an average diameter of about1.6nm. X-ray photoelectron spectroscopy (XPS) of the Ir4f7/2core level showed a main contribution from metallic Ir0after reduction at300℃with5vol.%H2/Ar. To clarify the origin of the enhanced LA conversion activity achieved by using Ir/CNT, H2temperature-programmed desorption (H2-TPD) measurements for Ir nanoparticles deposited on different supports were conducted. It is revealed that the H2desorption from the Ir/CNT occurred from lower temperatures and higher amounts than that from other catalysts, which implies that the adsorbed hydrogen species on the CNT surface could be more active for an reductive transformation. In view of the high activity of Ir/CNT catalyst, the reaction was conducted under very mild conditions. Amounts of intermediate hydroxyvaleric acid (HA) was produced when the reaction was conducted at50℃and1atm H2. However the intermediate HA is rather unstable and intra-molecular lactonization to GVL occurs easily as the reaction proceeded.Given the fact that an equimolar amount of FA apart from LA is also produced during the lignocellulosic biomass hydrolysis process, we studied the effect of FA on the catalytic conversion of LA. When we deliberately added an equimolar amount of FA to the hydrogenation system, a significantly retarded LA hydrogenation was observed. The yield of GVL further reduced with the increase of reaction temperature. Bearing in mind the notorious low CO tolerance of the platinum-group metals (PGM) toward LA hydrogenation, we suspect that the CO produced during FA decomposition at high reaction temperature is responsible for this phenomena. We then turn our attention to carrying out the reaction at lower temperatures but with elevated hydrogen pressures, enhanced catalysts amounts and prolonged reaction time to improve the reaction kinetics. To our delight the yield of GVL can be improved significantly by increasing the amount of catalyst or hydrogen pressure. FA has attracted much recent interest in the area of green and sustainable chemistry because of its potential as a safe and convenient hydrogen carrier. So the above mentioned art is benefit for producing biomass-derived FA.2. Reduction of LA to GVL with H2derived from catalytic decomposition of FAGiven the fact that an equimolar amount of FA apart from LA is also produced during the lignocellulosic biomass hydrolysis process, the development of new efficient methods for GVL production using formic acid as an in situ source of hydrogen is much needed. The success of this new route not only improves the atom economy of the process, but also avoids the energy-costly separation of LA from the mixture of LA and FA in aqueous solution. By using the Au/ZrO2catalyst the1:1aqueous mixture of LA and FA can be quantitatively converted to GVL. A phenomenon worthy of mention during the LA reduction over Au/ZrO2is the rapid pressure increase in the interior of the autoclave reactor from0.5MPa to a maximum value of approximately6MPa in the first half hour of the reaction. A significant FA decomposition leading to H2/CO2formation can be responsible for such an effect. Compared to palladium, platinum or ruthenium nanoparticles supported on zirconia as reference catalysts, gold is far superior to other noble metals for the reduction of LA with FA. These results can be confirmed with sole FA decomposition and LA hydrogenation with hydrogen in the presence of little CO. Bearing in mind the notorious low CO-tolerance of platinum group metals toward formic acid electro-oxidation, it is considered that the CO produced during FA decomposition may severely poison the Pt, Ru or Pd-catalyzed LA reduction using FA as the hydrogen source. Through acidic hydrolysis (catalyzed by0.5M H2SO4) of cellulose, we obtained an aqueous solution containing LA and FA. After partial neutralization with CaO and removal of insoluble solid by filtration, the aqueous mixture was transferred into an autoclave containing Au/ZrO2. Performing the reaction at150℃for8h produced GVL in97%yield (based on the conversion of LA).3. Conversion of LA, FA, amine into5-methyl-2-pyrrolidone on the basis of reductive amination methodology and reactive extraction of LA and FA into GVLBased on above results, we reported a one-pot conversion of LA and ammonia or primary amines into valuable and useful5-methyl-2-pyrrolidones that are currently based on fossil resources by using the Au/ZrO2-HCOOH-mediated reductive amination methodology. We found the formation of formamide during the reaction, however, the intermediate could be further converted into5-methyl-2-pyrrolidone with extension of the reaction time. A lot of GVL byproduct produced at high reaction temperature or diluted solution. A control experiment was conducted with GVL and amine in the presence of Au/ZO2catalyst, but no5-methyl-2-pyrrolidone was produced in the similar reaction condition. Based on the above results, we concluded the direct formation of5-methyl-2-pyrrolidone was a main route and didn’t pass through the formation of GVL.In the previous work, we have demonstrated that a highly active and robust catalyst based on gold deposited on acid-tolerant ZrO2can be used to convert an equimolar aqueous mixture of LA and FA into GVL in excellent yields. However, the production of GVL is complicated by the need to separate LA and FA from sulfuric acid with cost neutralization, as residual sulfur leads to low catalytic activity and deactivation with time-on-stream. The production of GVL with an efficient alcohol-mediated reactive extraction protocol has been proposed in order to facilitate the recovery of sulfuric acid. First, through an acidic hydrolysis (catalyzed by0.5M H2SO4) of cellulose, a mixture of LA and FA can be obtained. Then LA and FA can be converted to hydrophobic n-butyl levulinate (BL) and formate (BF) which separate spontaneously from sulfuric acid aqueous solution after n-butanol was added. The mixture of levulinic and formic esters can be directly converted to an aqueous solution of GVL and n-butanol over a single Au/ZrO2catalyst, in which H2in situ generated from BF is used for the reduction of BL to GVL. 4. Hydrogenolysis of GVL to1,4-PDO or2-MTHFRecently, Ru-based molecular catalyst system can selectively convert bio-derived LA into1,4-PDO or2-MTHF via intermediate formation of GVL, but the catalyst preparation process is complex, and need to add plenty of ligands and additives. We found the direct conversion of GVL into1,4-PDO or2-MTHF was realized by chemoselective hydrogenolysis catalyzed by a simple yet versatile copper-zirconia catalyst system. The Cu/ZrO2catalyst obtained by600℃-calcination (Cu/ZrO2-600) was used for hydrogenolysis of GVL to1,4-PDO. Firstly we studied the effect of copper loadings on the activity of GVL hydrogenolysis. An increase in Cu loading cause a significant improvement in the yield of1,4-PDO. The highest product yield was achieved when Cu loading was increased to30wt%. However, a further increase in the Cu loading to40wt%leads to a slight decrease in the desired product yield. By a careful correlation of the metallic Cu surface area data, it could be found that there is a good relationship between the metallic copper surface areas and the performance of the Cu/ZrO2-600catalysts with various Cu loadings. Studies on the effect of the reaction temperature at the same hydrogen pressure revealed that an obvious increase in the yield of the2-MTHF (ca.13%) when the reaction was conducted at240℃. However, the2-MTHF yield remained constant with further increasing the reaction temperature.In an attempt to improve the yield toward2-MTHF synthesis, subsequent studies were focused on the hydrogenolysis of GVL at240℃over a series of Cu/ZrO2catalysts with significantly modified acidic properties of the catalyst surface obtained by calcination in air at different temperatures in the range of300-700℃for4h. It was found that the Cu/ZrO2-400catalyst obtained by400℃-calcination can deliver a remarkable conversion of GVL to give2-MTHF in an excellent yield of ca.91%at240℃,6MPa H2within6h. In order to explore the reaction mechanism, the dehydrative cyclization of1,4-PDO to2-MTHF over the Cu/ZrO2-400catalyst under identical reaction conditions was possible, albeit at a slower rate than in direct carbonyl reduction. The Cu/ZrO2-400catalyst exhibits a higher abundance of weakly acid sites, as reflected from the significant desorption features appeared in the temperature region of200-400℃as seen in NH3-TPD experiment. This finding indicates that a synergistic cooperation between dispersed Cu and the acid sites of the catalyst surface is essential to facilitate the direct reduction of the carbonyl group in the GVL molecule or accelerating the subsequent dehydration of intermediate1,4-PDO to afford2-MTHF.