Sviluppo di un processo industriale per l'idrogenazione catalitica di zuccheri lignocellulosici per la produzione di precursori di bioPET

This PhD work has been focused on the study of the process of heterogeneous catalytic hydrogenation of a sugar stream, to produce a solution rich in sugar alcohols. This study is part of the development of a technology platform that can produce, from sugars extracted from lignocellulosic biomass, two chemicals of great industrial interest, ethylene glycol and 1,2-propylene glycol. This PhD work was done in apprenticeship in Biochemtex, R&D company of the Mossi & Ghisolfi group. From the synergy of the M&G group, leader in PET packaging, and Biochemtex, biomass R&D center, comes GREG, GReen Ethylene Glycol, the aim of which is to produce ethylene glycol, one of the main PET precursor, no more from oil, but from biomass extracted sugars. The catalytic hydrogenation of sugars to sugar alcohols is a known process in the literature and already industrialized for the production of sorbitol, xylitol and mannitol, which have found in recent years a number of applications, for example in the field of food industry, as alternative low caloric sweeteners. Nevertheless, being the overall current sugar alcohols production scale at about 1 Mt/y, the actual technology, based on batch reactors and problematic catalysts such as Raney nickel, would not be adapted to face a tenfold increase in productivity, which will be required for then obtaining amounts of ethylene glycol and 1,2-propylene glycol which may be in competition with the market of their equivalents from oil. Therefore, this research work has been dedicated to develop a new process for the hydrogenation of sugars, which is more suitable for continuous industrial production and high production capacity; in particular three main aspects of the reaction have been addressed: the catalyst reactivity, the chemical kinetics and the reactor design. With a batch system and a continuous trickle bed reactor, experimental tests were carried out to optimize the operating conditions and maximize the yields of the process. At first we used a solution of synthetic sugars and then the collected data have been verified with a purified sugar solution, confirming the feasibility of the process. Moreover, in order to obtain scale-up information, some experiments were carried out on the pilot plant. The selected reference catalyst is a 2wt% Ru/C, supplied by the British Johnson Matthey group, characterized by a high surface area, microporous morphology and an high dispersion degree of the metal active phase. The literature shows that the reaction of hydrogenation of xylose to xylitol with a catalyst based on a ruthenium follows first order kinetics, but experimental tests carried out in the course of this work have shown a different influence of the initial concentration of xylose. Consequently, tests were carried out in experimental kinetic regime and these have confirmed that the hydrogenation of xylose is not of first order and processing all the data collected, a reaction order of 0,3 was obtained. In addition, , the corresponding kinetic constants were estimated using the Arrhenius equation at four different temperatures (70, 85, 100 and 120° C) and the activation energy necessary for the reaction was calculated, obtaining a value of 34.1 kJ/mol, in agreement with literature evidences for the hydrogenation of different sugars. An extensive experimental campaign has been carried out, in both batch and continuous reactors, to investigate hydrogenation performances on a broad range of process conditions, using synthetic sugars and biomass-derived purified sugar streams. It was observed that, in the case of the purified sugars, a loss of conversion occurred after a certain period of time on flow, indicating that the catalyst was being deactivated. This phenomenon has been extensively investigated, to try to explain and prevent it. Consequently a detailed characterization of the exhausted catalyst was carried out and two other Ru/C catalysts with different morphological characteristics were tested. On the basis of this it was possible to conclude that the deactivation occurs by occlusion of the pores of the catalyst due to the accumulation of organic matter. This deactivation problem has been addressed in parallel, following three different paths. In the first one, it was chosen to test catalysts with different morphologies, in particular with an higher pore diameter. In the second one, a washing procedure was designed and tested, with the aim of restoring the initial catalyst activity. Finally, the influence of reaction conditions, in particular the temperature, on catalyst lifetime was thoroughly studied, in order to find an optimum set of temperature which could prolong it as much as possible. The experimental tests carried out with different catalyst carriers, such as zirconium oxide, titania and alumina, showed lower conversions of the carbohydrates present in the solution and have shown that these mesoporous catalysts with lower surface areas are less active in the hydrogenation of sugars to sugar alcohols. The washing procedure was tested in an experiment which, after about 200 hours at total conversion, showed loss of performances. To restore the total conversion for a further 70 hours was possible by washing the catalyst in situ with a 10%wt solution of sodium chloride under operating conditions and then flushing water at 200° C. Finally, the third way followed was to adapt the operating conditions for the use of purified sugar liquid, in particular as regards to the temperature. To investigate the influence of the presence of C6 sugars, tests were then carried out with mixtures of xylose and glucose, as compared to tests containing only a single sugar. Results have shown that, not only glucose has a lower reactivity than xylose, but when present in a mixture, it also has a competitive effect , probably due to a difference in the energy of absorption on the catalyst, which decreases C5 conversion. Because of this different reactivity, it has been observed that, although for synthetic sugars solutions temperatures of about 85-100°C are sufficient to obtain a total conversion, to hydrogenate completely sugar liquids from lignocellulosic biomass the temperature should be increased to about 130°C. From the comparison of the tests carried out in the laboratory and pilot plant scale with the same operating conditions and the same purified sugars liquid, it was observed that there is a significant difference in performance, in terms of catalyst lifetime. A possible explanation for this difference was found in the not optimized configuraiton of the pilot system, considering that the actual fluid dynamic behavior does not allow the liquid to interact with the catalyst with the desired contact time. With the aim of better understanding this phenomenon, a fluid dynamic study was carried out with empirical formulas found in the literature, at the specific operating conditions used in the system, and subsequently validated with experimental tests using a Plexiglas tube of the same size of the pilot reactor. Having the study revealed that the current use of the catalyst bed in the pilot plant was around 37%, the experimental study was then focused on the improvement of the wetting efficiency of the catalyst. The outcomes of the experimentation on the Plexiglass reactor were then implemented on the pilot plant, leading to a consistent improvement of catalyst lifetime. Simultaneously with these optimization activities, GREG process has been completely demonstrated at pilot plant scale, producing, starting with purified sugars from lignocellulosic biomass, a first batch of bio ethylene glycol, used in the synthesis of bioPET. Despite the characterization tests have not shown significant differences between the EG resulting from oil and the one from second generation sugars, there is still a slight difference in color in the finished PET, although better compared to a PET produced using EG first generation, obtained from ethanol produced by fermentation of sugars derived from sugar cane. During the process of molding and blowing, the addition of a dose of conventional toner has allowed to eliminate the residual coloring traces and to obtain a product equivalent to a commercial one.