Synthesis of supported catalysts of Co, Ni and Ru using supercritical Co2 for Biomass valorization reactions

Résumé

Nanostructured materials have been attracting increased attention for a wide variety of applications due to their superior properties compared to their bulk counterparts. In the literature, it can be found attempts to insert transition metals into mesoporous frameworks, generally by wet impregnation or co-precipitation showing various drawbacks such as an excessive use of solvents, high energy consumption or costly refining steps. In this way, the preparation of materials with supercritical fluids (SCFs) has been developed as a new research area of interest, and particularly in the field of nanoparticles, thanks to their ability to control the size and shape of the products changing its operating conditions. In this line of research, the aim of this PhD thesis is to make studies in order to develop an efficient process to synthesize different supported metal catalysts in supercritical medium. It is expected to obtain nanocomposites that can be used to enhance the selectivity for biomass conversion to produce compounds with high added value. In Chapter 1, a brief introduction on supercritical fluids, their properties and main uses in the synthesis of materials is presented; and some short notions about catalysis. Then, in Chapter 2, it has been decided to start with the production of cobalt-based catalysts due to its numerous applications and because there was not any previous work in which supercritical technique would be used for its synthesis. Cobaltocene (CoCp2) has been selected as organometallic precursor owing to its relatively high solubility in sc-CO2, and MCM-41 as support due to its sharp pore distribution, large surface and thermal stability. To choose the optimal parameters of deposition process, the adsorption isotherm of CoCp2 on MCM-41 has been determined, where it can be observed that this system presents low adsorbate-adsorbent interactions with an equilibrium adsorptive capacity around 65 mg/g according to the predictions estimated by the fittings. After that, it can be concluded that it is not enough simple exposure of a solution of the precursor over the support for its adsorption, being necessary to use the supercritical fluid reactive deposition (SCFRD) to promote the precipitation of metallic particles on the surface of support. This process has been optimized to find the combination of parameters that maximize the adsorption. At the end of this work, it was concluded that it is preferable to operate at 200 ºC and 14 MPa during one hour with the precursor and support distributed through the reactor to solve the problems caused by mass transfer limitations. Besides, it has proved that it is possible to achieve high cobalt loadings by means of consecutives depositions (up to 15 %wt in only 3 batches). Then, Co/MCM-41 samples were characterized by different technique and compared to other ones prepared by conventional synthesis methods, revealing that our catalysts showed the better metal dispersion on the surface support (~10%) and small particles diameter (between 1 to 9 nm), accessible even inside the pores by reactive molecules. Finally, an activity test was carried out. Catalysts were tested in the gasification of cellulose, increasing the selectivity for hydrogen production from 7% to 30% as a result of the promotion of tar, as it can be concluded with the fitting of the experimental data by a three-reaction model. To continue the work, it has been decided to check the versatility and robustness of the supercritical process with a different precursor-support system, although very similar, in Chapter 3. In this sense, nickel-based catalysts from nickelocene (NiCp2) on MCM-48, other member of the family of MS41 silica mesoporous support, have been synthesized by SCFRD. For this purpose, the adsorption isotherm of NiCp2 on MCM-48 has been extensively studied, revealing a completely different behavior. Experimental data have been adjusted by various models, being Toth equation which provided the best fit, and predicting a maximum value of adsorption of nickel around 130 mg/g at 70ºC and 14 MPa. In order to have a better understanding of the process, its kinetics has been modeled by pseudo-first and pseudo-second order, in which it can be observed that this adsorption requires times higher than 10 hours. This fact it could be attributed to some diffusional limitation. To corroborate this theory an intraparticle and film-diffusion models have been used to analyze the experimental data and determine the adsorption mechanism. Results resolved that the external mass transport is the rate-controlling step for this system. With this information, it has been decided to use the SCFRD with the same temperature and pressure established as optimal conditional in Chapter 2. Ni/MCM-48 samples with metallic loadings between 2-5 %wt have been synthesized. These catalysts have been also characterized, revealing the existence of nickel oxides homogeneously distributed on the surface of support and BET areas around 1000 m2/g. The employment of high pressure has not modified the original pore structure, unlike catalysts prepared by direct synthesis where the TEM pictures showed irregular surfaces and collapsed pores. Finally, the activity of catalysts were tested in the selective conversion of glucose to sorbitol, reaching conversion of 6%, selectivity higher than 90 % and yields around 5 % in 30 minutes, similar to catalysts synthesized by wet impregnation and improving substantially data obtained by catalysts prepared by direct synthesis. As conclusion, both in the first case of study about cobalt where the adsorption was practically non-existing, and in this last one about nickel in which the process is very slow, the SCFRD has proven to be a viable technique. On the other hand, in Chapter 4, a continuous process for the synthesis of catalysts based on the supercritical fluid adsorption has been developed. With this aim, it was necessary to select an appropriate metal-support system, and start its study from the beginning. First, the solubility of ruthenocene, RuCp2, has been determined at 60 ºC and different pressures by cloud point technique in a visual cell. Experimental data have been fitted by Chrastil and Peng-Robinson with god agreement. Results are consistent with data found in literature. Then, the adsorption of RuCp2 on two supports with different structure, one mesoporous (MCM-48) and other microporous (activated carbon) has been studied. Both cases have shown good results where the metallic loading in the support increased with the supercritical phase concentration, characteristic of monolayer deposition. Experiments conducted at different pressures (9, 11, 14, 17 and 20 MPa) have been revealed how the process is favored at low pressures. Following a similar analysis carried out in the previous chapter, experimental data have been adjusted by different models, being the three-parameter equations such as Redlich-Peterson or BET correlations which provide the best fit. The study seems to reveal as there is a direct relationship between the controlling step of the process and its kinetics. Although both mechanism have complex mechanism with different stages involved, in the adsorption on activated carbon the controlling step is mainly the external transport, with a value of saturation limit of 80 mg/g after 10 hours, similar to nickel-based catalysts. On the other hand, the adsorption on MCM-48 is largely governed by pore diffusion, reaching loading up to 15 mg/g in shorter times, around 4 hours. Thanks to this information, ruthenium-based catalysts have been satisfactorily synthesized by supercritical fluid deposition in a continuous process, at 60 ºC and 11 MPa. Different values of CO2 flows have been tested (15, 30 and 45 mg/g) to avoid the problems owing to transfer limitations and fitted by first order model with acceptable average errors. Catalysts have been characterized and proved in the hydrolysis of cellobiose for glucose production. Ru/MCM-48 with metallic loading of 4 %wt have shown the highest yield, improving the process without catalysts from 18% to 48%, with selectivity around 90%. Besides, experiments with different loadings have been carried out, revealing that a higher amount of metal produce the collapse of pores and avoiding the contact with the active centers.