Kinetics for Cobalt Fischer-Tropsch Synthesis; A Multi-Dimensional Task

Kinetics for Cobalt Fischer-Tropsch Synthesis; a multi-dimensional task

Magne Hillestad, ^a Mohammad Ostadi, ^a and Erling Rytter ^a,b

^a Department of Chemical Engineering, Norwegian University of Science and Technology (NTNU), N-7491 Trondheim

^b SINTEF Industry, N-7465 Trondheim, Norway

1.    Introduction

In order to achieve an optimized and realistic design of a Gas-to-Liquid (GTL) or Biomass-to-liquid (BTL) plant based on cobalt Fischer-Tropsch synthesis (FTS), it is necessary to have available a kinetic model for FTS that correctly capture major variations in reaction rates and selectivities subject to changes in process parameters. Unfortunately, the kinetic responses vary to a large degree with the catalyst formulation, in particular the type of support material for cobalt crystallites. Water is a key component in the reactions, and the activity response changes from positive to negative depending on the support, even within one family of supports like γ-alumina.^[1] Add to this experimental data with restricted variations in process parameters, data obtained in different reactor types, a multitude of approaches for formulating kinetics, lack of agreement on the FTS mechanism, and several parallel or interlinked reactions. The latter comprise formation of long chained paraffins, the associated chain-propagation probability, formation of olefins, and methane as an apparent separate product. The task becomes multi-dimensional with several possible outcomes.

Here we present comparison of thirty published kinetic rate models dating back to 1949, comprising twenty-four explicit rate expressions. Some of the models were used in optimization of a BtL plant. Further, the multiple-dimension of formulating a good and useful kinetic expression for FTS has prompted us to revisit both the experimental basis, formulation of useful and reasonable expressions based on understanding of the FT reaction,^[2] and, provisionally, implementing new mechanistic based Langmuir-Hinshelwood formulations.^[3] An important part of this effort is flexible handling of water responses and catalyst formulations.

2.   Comparison of published kinetic expressions

Some of the reported kinetic models give a detailed description of formation rates of individual components, while most models only describe the overall consumption rate of CO. Twelve of them were analyzed based on different criteria, such as their behavior at high conversions, high water partial pressure, sensitivity to added water, selectivity to C₅₊ products, etc.^[4] The rate models were implemented in a plug flow reactor model quite accurately representing both fixed-bed and microchannel reactors. The main objective was to see the kinetic effects with changing composition along the reactor. In order to predict product distribution, we proposed our own chain growth model and fitted parameters to published experimental data.^[5] The chain growth model includes the effect of water and predicts the C₅₊ and methane selectivities quite well.

Figure 1. Activity as a function of CO conversion based on 12 different rate expressions; see ref. 4. a) Under-stoichiometric hydrogen feed; H₂/CO=1.5. b) Over-stoichiometric hydrogen feed; H₂/CO=2.5 .

An example of the results is shown in Figure 1. The modelled rates span large variations depending on the rate expression with particular large discrepancies at high conversions. The spread is particularly large for feed gas with excess hydrogen. On the other hand, it is gratifying that the differences are not that large, except for one expression, for the core conversion range between 40 and 60% and with the industrially preffered under-stoichiometric feed. However, pushing the conversion to 70 or 75%, relevant for multichannel reactors, causes concern and necessitates careful selection of rate expression.

3.    Optimizing a BTL plant depending on FT kinetics

This part of the study is concerned with optimization of a BtL plant. The basic configuration, shown in Figure 2, is based on biomass gasification in an entrained flow gasifier, supply of additional water from a high temperature electrolysis unit (SOEC), reverse water-gas-shift (WGS) reaction to convert CO₂ to CO, and multistage FTS reactors.^[6] By adding the extra hydrogen, more than 90% carbon efficiency from biomass to products are achieved.

Figure 2. Block-flow diagram of BtL concept.

Several kinetic expressions from the study mentioned above are compared and used in process optimization of the plant. Fixing the sizes of the FT-reactors results in variations in total hydrocarbon productivity and the optimal H₂/CO feed ratio. Still, the main conclusion stands; it is profitable to add hydrogen from an external source, preferable made from renewable electricity, due to increase in carbon efficiency.

5.    References

---

[1] E. Rytter, Ø. Borg, N. E. Tsakoumis and A. Holmen, Water as key to activity and selectivity in Co Fischer-Tropsch synthesis; γ-alumina based structure-performance relationships, J. Catal., 365 (2018) 334-343.

[2] E. A. Blekkan, M. Hillestad, E. Rytter, L. Gavrilovic and co-workers, work in progress.

[3] E. Rytter and A. Holmen, Consorted Vinylene Mechanism for Cobalt Fischer-Tropsch Synthesis Encompassing Water or Hydroxyl Assisted CO-activation, Topics Catal., 61 (2018) 1024-1034.

[4] M. Ostadi, E. Rytter and M. Hillestad, Evaluation of kinetic models for Fischer-Tropsch cobalt catalysts in a plug flow reactor, Chem. Eng. Res. Design, 114 (2016) 236-246.

[5] Yang, J., B, S., Myrstad, R., Venvik, H.J., Pfeifer, P., Holmen, A., 2016. Fischer–Tropsch Synthesis on Co-based Catalysts in a Microchannel Reactor. Effect of Temperature and Pressure on Selectivity and Stability. Catalysts and Catalysis: advances and Applications, vol. 223. CRC Press, pp. 259–266.

[6] M. Hillestad, M. Ostadi, G. d. Alamo Serrano, E. Rytter, B. Austbø, J. G. Pharoah, O. S. Burheim, Increasing carbon efficiency of the biomass to liquid process with hydrogen from renewable power, Fuel, 234 (2018) 1431-1451.