(135c) Selective Hydrogenation of Light Hydrocarbons Products of Flash Hydropyrolysis of Grasses Pastures Using a Fluidized Bed and Pd Ce Ni Eggshell Catalyst.

Roberto Galiasso Tailleur^a,b,c

^aSimon Bolivar University, Baruta, Caracas Venezuela and ^bHypro-Consultant, 4250 Corrine Dr Suite 204, Orlando Florida 32814

^croberto.galiasso@hyproconsultant .com

Extended Abstract

Hydrogenation of the fraction C₁-355°C, produced during the flash hydropyrolysis of grasses pastures was performed in a fluidized bed reactor using a thio-resistant PdNiCe on silica eggshell catalyst. The catalyst, in bids of 80 micron of diameter, contains 0.1% of Pd, 0.2% of Ce and 0.3% of Ni on surface. Metals were incorporated on surface by impregnating the gel precursor of the support and then calcining it at 480°C. Before used the catalyst was reduced in hydrogen at the same temperature. The effects of temperature and pressure in apparent rates of olefins and diolefins hydrogenation and catalyst deactivation were measured in 25 short tests and then in 7 day pilot plant continuous operation. The feed was obtained from a high-pressure high-temperature separator (HPHTS) of the inverted hydrocyclone reactor [1] pilot plant that process perennials forage. The gas phase, obtained by the top of HPHTS at 12MPa and 593 K, contains H₂, H₂S, water, NH₃ and C₁-355^oC hydrocarbons; it was cooled to either 633, 643 or 653K and depressurized to either 11, 9 or 7MPa before being introduced in a fluidized bed reactor (HDT) to be hydrogenated. Catalyst was added and withdrawn from the bed continuously to maintain constant the activity. The heavy liquid fraction of HPHTS can be used, among other possibility, as feed in the Delayed Coking (DC) unit. The composition of the light-hydrocarbons and those of the hydrogenated products are shown as an example in Table 1. The feed to the HDT reactor contains olefins and diolefins with the following distributions: 40 and 36% in the C₄-C₅, 43 and 48% in C₆-C₁₀ and 17 and 12 % in the C₁₀-355 °C fraction, respectively. The HDT product obtained at 9 KPa, 543 K, 0.2 and 0.4h of gas and solid residence time shown that there is more than 90 and 56% of diolefins and olefins conversion. Most of none-aromatics sulfur (mercaptans, and other thiols), oxygen (carboxilic acid, ester, and alcohols) nitrogen content hydrocarbons were reduced. Thus, product acid content was reduced as well as the amount of solid formed during the accelerate storage test (one week at 55°C under nitrogen) reported in Table 1. The spent catalyst contains an average of 1.1 wt % of coke. Heteroatoms containing hydrocarbons in aromatic structure are not modified by the HDT at this condition. It is used to remove the most reactive olefins and diolefins that accelerate the HDS catalyst deactivation.

Table 3 shows the results of a continuous operation for 8 days at different operating conditions. Each condition was maintain for one day. Notice the effect of temperature, gas residence time and hydrogen partial pressure in olefins and diolefins conversion, as well as the catalyst throughput used to maintain the activity. The mass transfer calculation shows that hydrogenation of diolefins and olefins are not affected by diffusion from gas to external surface of the eggshell catalyst. The reaction produces additional gases, hydrogenated naphtha-diesel fraction (C₅-355°C) and coke. The hydrogenated product needs to be treated in a conventional HDS (hydrotreater) to adjust further the properties and fulfill the fuel specification of the fractions.

Coke is accumulated on catalyst in multiple layers on active sites and on silica support; coke is withdrawn with the spent catalyst and then it was regenerated by combustion; particles have a carbon content distribution equivalent to a perfect mixed flow function (E(θ):1.1 CSTR). Higher the temperature and lower the pressure and the solid throughput, higher the amount of coke accumulated on surface. Nevertheless, not all the coke is deposited on active sites. The probability that coke grows on top of acid sites (Po) increase with temperature between 0.16 to 0.20 (Table 2) but is not affected by pressure. Coke dispersion was measured by high resolution electron microscopy and XPS analyses of the surface. The use of silica bids instead of catalyst at the above mentioned three operating temperature shows a higher coke deposition of coke (1.6, 2.3 and 2.6 wt. %) than in presence of catalyst, but there are the same distribution of particle size on surface (between 12-40 microns diameter). Average coke content at the steady state operation with PdCeNi catalyst is 0.82, 1.1 and 1.4 wt. % on the solid withdrawn respectively at 533, 543 and 553 K (9MPa, 0.2 and 0.4h of gas and solid residence times). The solid flowrate was changed according to temperature, gas residence time and pressure of operation used in the plant. The solid throughput employed at different operating conditions was calculated using a model of reaction and deactivation obtained previously [2]. The model consider that deactivation of hydrogenation active sites is dependent on olefins diolefins and hydrogen partial pressure (equation (1)); the apparent kinetic rates are based on olefins and diolefins disappearances with deactivation. Olefins and Diolefins are hydrogenated on the solid (adsorbed species), but they are polymerized in gas phase producing nanoparticles of coke that deposit on catalyst; coke rate of built-up depend upon olefins and diolefin olefin and hydrogen in gas phase (see equation in (1); catalyst activity (a) depend upon amount of coke formed and the probability of deposition on active sites at the external surface (0.09 m²/g); Rate of reaction and deactivation were solved using a Runge-Kutta and Genetic Algorithm to obtain the values of the rates constants; they are needed to predict catalyst throughput at different operating conditions.

The results obtained during 7 days of continuous operation, displayed in Table 3, confirmed the effects of variables in activity and the stability of the operation. There, the effects of variables in olefins diolefins, sulfur and hydrogen consumption can be seen in the rows. Notice that the system achieved the same level of conversion at the beginning and at the end of tests thank to the use of adequate throughput of catalyst mentioned in the table. The information obtained was employed to confirm the values of the activation energies that are reported in Table2. The table shows that activation energy for catalyst deactivation (E_d) is 21 and 35% higher than those of hydrogenation of diolefins (E_dol) and olefins (E_ol). Deactivation occurred by coke coverage of active sites. The XPS analyses of coke and Ni content on surface indicated the metal dispersion decrease with the logarithm of coke content. HDT stage is operate at low temperature to minimize the gas phase polymerization of the most reactive olefins and diolefins. The hydrogen consumption is low at this conditions and the catalyst can be regenerated.

[1]Galiasso Tailleur R. Gonzalez Y, Lucena M.;"New invertd cyclone reactor for flash pyrolysis (2014) Catalysis Todays 220-222,186-197.

[2]Galiasso Tailleur R. "Hydrogenation and desulfurization in gas phase of light hydrocabons from hydrocracking and delayed coking. (2019) Catalyst deactivation. Chem. Eng. Sci. 210,3 115-195

r_Ol(mol/gh)=k_Ol.P_H2^0.3.P_Ol^0.2.a_f; r_Diol (mol/gh)=k_Diol^.P_H2^0.3.P_Diol^0.3.a_f ;r_coke (wt%/h)=k_cokeP_Ol^0.3.P_Diol^0.28.(0.28+PH2)^-0.12; r_{d ^(a/day)}=k_d^.Wcoke^.Pr.a ; t_s(Lh/L):(1-a_f)/(k_d.a_f); a_f=1/(1+r_d) (1)