Multiphase reaction studies in stirred tank and fixed bed reactors

Shinbeom Lee, Purdue University

Abstract

A biphasic stirred tank reactor and a trickle bed reactor were studied to understand complex multiphase reactor behavior arising from mass transfer effects on reactions and to improve modeling accuracy for rational design and optimization. For the first part involving a stirred tank reactor, the intrinsic reaction rate of n-butyraldehyde aldol condensation was obtained in the industrially relevant range 110–150 °C and 0.76–1.9 M NaOH, which is in the mass transfer regime dominated by reaction in the film. A stirred cell was used to obtain stable interface between the organic and aqueous phases. The mass transfer regime was confirmed by plateau region experiments and calculations of mass transfer. As a result, considering nBAL solubility and diffusivity, the rate was found to be 1st order in both nBAL and NaOH concentrations, along with 13.5±0.4 kcal/mol activation energy. The kinetic parameter sensitivity using different models for solubility, diffusivity and salt effect was also studied. This work demonstrates that, using penetration theory, it is possible to determine intrinsic reaction kinetics in the mass transfer regime, governed by reaction in the film. Following the first step, reactor modeling for n-butyraldehyde aldol condensation was investigated under the industrially relevant conditions. The interfacial area in the reactor was directly measured using a borescope system under appropriate temperature, NaOH concentration and rpm conditions. To estimate the interfacial area, a semi-empirical correlation was developed, which provided good estimates within ±15% error. The reactor model based on the two-film theory was developed, combining the interfacial area and intrinsic reaction kinetics reported above. The model was verified by reaction experiments in the range 0.05–1.9M NaOH, 80–130 °C and 600–1000 rpm, similar to the industrial conditions. The prediction errors of the reactor model, combining the interfacial area from direct measurements and the correlation were ±8% and ±15%, respectively, suggesting that the model accuracy may be improved with better interfacial area estimation. For the study of a trickle bed reactor, intrinsic kinetics and internal diffusion effects using various support sizes were investigated for acetophenone hydrogenation. The 1% Rh/Al2O3 catalyst was selected by catalyst screening tests using different noble metals and supports in a slurry reactor. Intrinsic reaction kinetic modeling with the Langmuir-Hinshelwood mechanism was conducted from experiments at 60–100 °C, 1.1–4.1 MPa PH_2 and 0.04–0.4 M CAP.o using powder catalysts. The selected kinetic model included dissociative and non-competitive hydrogen adsorption, along with saturated active sites for organic species, and surface reaction as the rate determining step. With the obtained intrinsic reaction kinetics, internal diffusion effects were investigated using two catalyst particle sizes and diffusion-reaction models. The properties of the egg-shell type catalyst particles, including metal dispersion, were characterized and utilized in the models. The predictions of the models developed in this work correspond well with the experimental results, explaining the effects of internal diffusion inside catalyst particles on reaction rates and selectivity. In a trickle bed reactor, flow regime effect and reactor modeling studies were conducted for acetophenone hydrogenation on 1% Rh/Al2O 3 catalyst, a relatively high pressure and complex reaction scheme typical for pharmaceutical applications. The reactor consisted of a 7.1 mm ID stainless steel tube with 0.5 mm catalyst spheres. From hydrodynamic tests, trickle and bubbly flow regimes were confirmed visually with regime map developed for different gas/liquid, tube/particle materials, pressure and temperature. The operating conditions for each regime were identified using pressure drop fluctuations for the opaque stainless steel reactor. The beneficial effect of bubbly flow on reaction rate was confirmed experimentally in 0.02–0.19 m/s and 2.5–12 mm/s for gas and liquid superficial velocity ranges, respectively under 80–100 °C, 11–26 bar and 0.04–0.6 M CAP.o conditions. The effects of partial wetting and liquid limited reaction were suggested from studies involving gas flow rate, temperature and pressure variation. The reactor model including external/internal mass transfer along with the flow regime effects was developed using an adjustable parameter for partial wetting and flow regime effects. With fitted parameters using a part of the experiments, the model provided good predictions (R2 >95%) for all experiments. The combined experimental and modeling approaches followed in the present work are good examples to demonstrate the effects of mass transfer on reactor performance. This thesis will help to improve the modeling accuracy for design and scale up with fundamental understanding of multiphase reactors.

Degree

Ph.D.

Advisors

Varma, Purdue University.

Subject Area

Chemical engineering

COinS