Graduation Year

2024

Document Type

Dissertation

Degree

Ph.D.

Degree Name

Doctor of Philosophy (Ph.D.)

Degree Granting Department

Chemical, Biological and Materials Engineering

Major Professor

John Kuhn, Ph.D.

Co-Major Professor

Babu Joseph, Ph.D.

Committee Member

David Simmons, Ph.D.

Committee Member

Andre Tejada, Ph.D.

Committee Member

George Philippidis, Ph.D.

Keywords

Reactors, Fuels, Zeolites, Diffusion, Optimization

Abstract

Despite rigorous study and advancement in fields of green energy, the world's transportation industry remains poised to continue reliance on liquid hydrocarbon fuels. To date, the primary source of diesel and jet fuels is the petroleum industry; however, continuing trends in rising greenhouse gas emissions present a need for study into processes for the production of synthetic hydrocarbon fuels. The most promising technologies to meet this goal are gas-to-liquid (GtL), biomass-to-liquid (BtL) and coal-to-liquid (CtL). The commonality of these technologies is that they utilize a feedstock to produce carbon monoxide and hydrogen gas (syngas) which becomes the building blocks of hydrocarbon products through Fischer Tropsch Synthesis (FTS). The primary challenges facing the application of FTS technology lie in the source of syngas. BtL processes are often inefficient and unsustainable and CtL results in high emissions rendering it unsuitable for carbon neutral and reduced emission applications. GtL processes have shown significant promise but are economically limited due to economies of scale. The reforming of methane is the most common source of syngas for GtL systems, but the process requires high temperatures (~800°C) and low to moderate pressures. When paired with the moderate conditions that are optimal for FTS (250°C, high pressure) maintaining energy efficiency, and thus profitability, becomes challenging. An approach to overcome this is to operate both reactions under similar temperature and pressure conditions; however, this method introduces new considerations. Methane reforming at low temperatures results in reduced syngas yield and high temperature FTS thermodynamically favors methane and light alkanes as products. Furthermore, the design of a novel intensified single-reactor system is an expensive and time-consuming process. While experiments and catalyst development is ongoing in our lab, the challenge of fully investigating the behaviors and interactions of these two reactions at abnormal conditions is required to guide experiments.

In this dissertation, a model of an intensified methane reforming and subsequent FTS reactor system was developed. The objectives of this study are threefold. Firstly, probe reaction behaviors of low temperature methane bi-reforming and high temperature FTS. This includes the dependence of product selectivity and activity on reactor parameters at these suboptimal conditions. Second, an investigation into the role of a zeolite layer on the diffusion of hydrocarbons. This includes evaluating the capacity of a zeolite to protect a reforming catalyst from heavy hydrocarbon products, determination of the optimum zeolite thickness to do so and the evaluation of the inter-molecular forces at play within the zeolite. The final goal is the study of a completed model for the intensified reactor. Evaluation of the model will improve understanding of the interplay between reforming and FTS in a low pressure, single reactor system. Furthermore, ideal operating parameters, feed composition and catalyst arrangement can be identified with the objective of maximizing product. The design of the model is broadly broken into three main components, namely the development of a low temperature methane reforming model, a high temperature FTS model and the unification of the two into a unified, single reactor process. Application of the unified model enabled the assessment of critical interactions between the two reactions under suboptimal conditions. This includes the impact of reactor operating parameters, feed composition, and arrangement/layering of the two catalysts (hereafter referred to as bed configurations). An additional secondary study was carried out to investigate the diffusion behaviors of mixed hydrocarbon species in an MFI zeolite. Zeolites as size-selective barriers have been proposed as a solution to scenarios in which FTS products may be broken down by the reforming catalyst. It was hypothesized that a zeolite may limit the diffusion of heavier products while leaving reactants such as methane relatively unhindered. To address this, a reactor scale zeolite diffusion model was developed to assess the effect of zeolite shell layer thickness on reactor performance. This study was performed in conjunction with the development of the reforming model.

The methane reforming model was developed on COMSOL utilizing kinetic models developed from literature for both the steam and dry reforming of methane. Steam reforming kinetics were proposed for a nickel reforming catalyst. Both models were applied to reproduce experimental data for a variety of temperatures, pressures and feed compositions to ensure accuracy of the initial model. Following initial verification, a fit was performed to experimental data from literature for bi-reforming of biogas at low temperatures (400-550°C). The fit to the testing data shows reasonable agreement with the experiment and system behavior follows known behaviors (thermodynamic effects of temperature and pressure on equilibrium and rate). Model scale-up was performed for implementation of heat transfer effects and comparison with available literature for industrial reforming. This was done due to the lack of thermal gradients at bench-scale. Mass transfer was incorporated into the model assuming gas-phase, bulk diffusion in line with the Fuller expression and Knudsen diffusion was assumed for diffusion in catalyst pores. For diffusion within the zeolite shell, configurational diffusion theory was used for the MFI-type zeolite, silicalite-1. The effect of zeolite shell thickness was investigated by comparison of the reforming of a variety of hydrocarbons including methane, propane, butane, heptane and toluene. Using a optimization criteria consisting of the rate of change in hydrocarbon conversion relative to methane normalized by the methane conversion, the point at which the relative conversion of a hydrocarbon relative to methane is greatest could be identified. From the analysis, it was observed that toluene and heptane were strongly affected by even a thin zeolite shell 10 nm in thickness. Thicker optimum shell thicknesses of 5-10 μm was identified for propane and butane.

An in-depth analysis of zeolite diffusion was conducted using a Kinetic Monte Carlo simulation for mixtures of C1-C4 hydrocarbons and benzene. The model considers a flexible zeolite lattice for the purposes of calculating the energy barrier of diffusion and assumes an activated diffusion process. The model was augmented through inclusion of Lennard-Jones type interactions between diffusing molecules and considers cases in which diffusing molecule create blockages in the lattice, interact with one another and potentially pass on another in the larger intersections of the zeolite. The model achieves reasonable agreement with the available experimental data for single and two-component diffusion. The model was then applied to isolate the effects of blockages, interactions and passing on the prediction of the overall diffusion coefficient. For mixtures of large and small molecules such as benzene and methane, lattice blockages had a significant effect on diffusion. For mixtures of moderate-light hydrocarbons blockages did not contribute to the overall process as much however molecular interactions could have beneficial effects on diffusion, especially on the larger of the considered species while the smaller molecule exhibited a slowing behavior. Diffusion coefficients for binary and tertiary mixtures were predicted.

An analogous model for FTS was designed in COMSOL for high temperature reaction. Kinetics for FTS and water-gas-shift on an iron catalyst were identified from literature and adapted to include corrections for methanation and reduced ethene selectivity (commonly observed behaviors in experimental FTS). The mass transfer of gases in liquid filled pores is estimated using empirical correlations. The kinetics were fit to experimental data for a high temperature FTS catalyst developed in our group lab. The fit achieved good agreement to the test data for a temperature range of 360-405°C and syngas ratio of 1-3. The model was applied in a sensitivity study observing the effect of temperature, pressure, syngas ratio and carbon monoxide conversion on the FTS product selectivity for a high temperature range. As anticipated, increasing temperature was strongly associated with a lighter product. The syngas ratio was also observed to control the product distribution. Notably, the effect of pressure and syngas ratio were coupled. The selectivity to heavy products was most sensitive to changes in syngas ratio at high pressures (10-20 bar) but did not change significantly at low pressure. It was also noted that the product distribution was fairly stable over a CO conversion range of 10-95%.

Finally, the two separate models for reforming and FTS were combined into a single intensified reactor. The model considered bi-reforming of methane at low temperature (650-750°C) and FTS at high temperature (400°C), both reactors are operated at the same temperature (1-10 bar). The final model incorporates both single phase and two-phase mass transfer as well as thermal control through separate electric furnaces for each reactor. The geometry of the system is based on a bench scale unit used in our lab for experiments. Application of the model yielded a close match to the experimental mass yield under the same conditions. The validated model was used to study the influence of process conditions, feed composition and bed configurations on the reactor performance measured through mass yield and product distribution. A maximum mass yield of C2+ products of 8.9% was obtained at a feed composition of 40% methane, 30% carbon dioxide and 30% water. Increasing the pressure resulted in a heavier product, however the losses in reforming performance lowered the methane and carbon dioxide conversion by 10% and 5% respectively. This ultimately lowered the mass yield of C2+ products from 8.9% to ~7.9%. Thus, a pressure of 5 bar was recommended to properly balance the two reactions. Despite simplifying assumptions favoring the configuration, only minor increases in mass yield were noted from the four-bed configuration. The two-bed process achieved a C2+ mass yield of 8.9% while the four-bed system reached a mark of 10.8%. This indicates that a two-bed system would achieve comparable results with significantly less complexity. An autothermal reactor incorporating a narrow temperature range (~200°C), multiple alternating catalyst beds and a ZSM-5 zeolite coating was also considered. Depending on the temperature range considered, the configuration was able to achieve a mass yield of 8.5% within 8-14 catalyst beds. Introduction of the encapsulated reforming catalyst had the expected result of reducing the reforming of FTS products, however, reduced methane conversion as a result of mass transfer limitations and blocked zeolite pores negated any improvements in performance and resulted in a comparable mass yield relative to the uncoated catalyst. Incorporation of the three key aspects of this study; low-temperature bi-reforming of methane, high temperature FTS and zeolite encapsulation of a reforming catalyst, presents a tremendous number of design considerations for reducing the energy demand of GtL processes while maintaining an appreciable mass yield. The reactor configurations screened here represent ripe opportunities for focused experimental study to improve the economics of GtL.

Download

Share

COinS