Kinetic modeling of liquid-phase hydrogenation reactions

Abstract

Kinetic models are an essential part of modern computer simulation based process design. The goal of the work presented here was to develop models for two types of industrially important hydrogenation reactions, namely the hydrogenation of monoand diaromatic compounds and the hydrogenation of isooctenes. The studied reactions are important in the production of new, environmentally friendly fuels. Saturation of aromatics is needed to limit the undesired particle emissions in the exhaust gases of diesel engines, and isooctane could be used to replace methyl-tert-butyl ether (MTBE) in gasoline as an octane rating increasing component.

The hydrogenation of aromatics was studied on a commercial Ni/Al2O3 catalyst using three model compounds: toluene, 1,2,3,4-tetrahydronaphthalene (tetralin), and naphthalene. The results showed that the first ring of the diaromatic compound (naphthalene) is considerably more reactive than the second, whereas the hydrogenation rate of the monoaromatic compounds (toluene and tetralin) is only slightly affected by the structure of the substituent. In addition, an inhibition effect induced by competitive adsorption was observed in the hydrogenation of multicomponent mixtures. The most reactive compound adsorbs most strongly on the catalyst and inhibits the other reactions. Hydrogenation was assumed to proceed by a mechanism of stepwise addition of dissociatively adsorbed hydrogen. Langmuir–Hinshelwood type rate equations were able to describe the reaction kinetics successfully, including the inhibition effect. The estimated adsorption parameters in the mixtures increased with reactivity (Ktoluene = 1.0 × 10−3 m3/mol, Ktetralin = 4.4 × 10−3 m3/mol, Knaphthalene = 7.8 × 10−3 m3/mol). Additionally, the estimated activation energies were in the physically meaningful range of 26-59 kJ/mol.

The hydrogenation of the isooctenes 2,4,4-trimethyl-1-pentene (TMP-1) and 2,4,4-trimethyl-2-pentene (TMP-2) to "isooctane" (IO; 2,2,4-trimethylpentane) was examined on commercial Ni/Al2O3, Co/SiO2, and Pt/Al2O3 catalysts. Qualitatively, the hydrogenation proceeded in the same way on the different catalysts in that TMP-1 (terminal double bond) was more reactive than TMP-2 (internal double bond), isooctane was the sole product, and double bond isomerization did not play an important role under the conditions used. Kinetic models were formulated on the basis of the two-step Horiuti–Polanyi mechanism, assuming rate limitation by the first hydrogen insertion. The difference in the activities (Ni > Co > Pt) of the three catalysts was concluded to be due to the number of active sites because turnover frequencies (TOFs) were of the same order of magnitude. However, in some features, Pt was found to deviate from Ni and Co in the hydrogenation of TMP-1 and TMP-2. Activation energies were higher (Eapp,TMP-1 = 49 kJ/mol on Pt and 34-35 kJ/mol on Ni and Co; Eapp,TMP-2 = 65 kJ/mol on Pt and 43-49 kJ/mol on Ni and Co) and hydrogen adsorption equilibrium constants were larger by two orders of magnitude (KH = 38 × 10−4 m3/mol on Pt, 0.16 × 10−4 m3/mol on Ni, and 0.30 × 10−4 m3/mol on Co). In addition, catalyst deactivation through the formation of carbonaceous deposits was considerably faster on Pt.

The kinetic equations developed in this work are applicable as such in reactor design because mass transfer, hydrogen solubility, and solvent effect were taken into account in the parameter optimization. In the hydrogenation of aromatics, valuable information was obtained on how to describe hydrogenation reactions in multicomponent mixtures, such as real diesel fractions. The results from the hydrogenation of TMP-1 and TMP-2 provide information that can be applied to the selection of an optimal catalyst material as well as in the design and optimization of industrial-scale reactors.