Solar Thermal Hydrogen Production: A Dynamic Reactor Model for the Hercynite Cycle

Abstract

The focal point of this thesis is the production of hydrogen. Hydrogen is classified as a secondary energy source and must, analogous to electricity, be efficiently produced from some primary energy source. This primary energy source has traditionally been fossil fuels. Steam methane reforming and gasification of coal are the most common routes for hydrogen production today. This thesis models and evaluates an alternative to fossil based hydrogen production, namely a solar thermal water splitting cycle.

A solar thermal water splitting cycle is a processes where a metal oxide is alternately reduced and oxidized in a cyclic fashion using solar energy and water. In the reduction step the metal oxide is thermally reduced by solar radiation at a temperature TR and oxygen gas is evolved. A re-oxidation of the reduced material is performed at a temperature TOX in a subsequent step where water is introduced to the system. Water will decompose on the metal surface and deposit oxygen while hydrogen gas is evolved.

The goal of this thesis was to model a proposed solar thermal reactor design for the so-called doped hercynite cycle and to evaluate the energy efficiency of the modelled system. There has recently been discussions in the solar thermal hydrogen production community regarding the optimal temperature of separation between the oxidation and reduction steps. The optimal temperature of separation can be defined as Topt = TR - TOX, where TR TOX and represents the most energetically efficient way of producing hydrogen. Solar thermal processes have traditionally been driven at constant pressure. However, it has recently been claimed that the optimal temperature can in fact be zero, i.e. isothermal operation, if a pressure swing is also imposed on the reaction system. It was desirable to contribute to this debate using modelling results.

The proposed solar thermal hydrogen reactor can either be a stand-alone reactor or it can, as current research at NTNU suggest, be implemented in a Fischer-Tropsch (FT) plant. The effect and feasibility of a solar thermal hercynite reactor implementation in an FT plant was also considered.

A dynamic, one-dimensional, pseudo-homogeneous reactor model was constructed and evaluated for the thermochemical production of hydrogen via the hercynite cycle. It was found that isothermal pressure swing operation was less energy efficient than combined temperature and pressure swing operation. The optimal temperature of separation based on maximum cavity receiver temperature (1623.15 K) was Topt = 63.15 K with a reduction system pressure of 0.21 bar and a oxidation system pressure of 5 bar. The oxidation temperature was Tox = 1560 K.

Based on energy efficiency requirements found in relevant literature, a target value for solar thermal reactor efficiency was a 20% solar energy-tohydrogen conversion efficiency. The highest calculated energy efficiency for a stand-alone hydrogen reactor was 9.7%. For a solar thermal Fischer-Tropsch (FT) plant coupled reactor the efficiency was estimated to be 11.7%. For every evaluated temperature reactor implementation in an FT plant increased total energy efficiency by 2 - 4%.

It was concluded that the designed hercynite cycle reactor would probably not be able to reach the 20% energy target based on the provided values. However, a rigorous optimization was not within the scope of the thesis; system optimization is a multivariable problem that is worthy of a thesis on its own. Better efficiencies and more optimistic predictions might result from a more rigorous system optimization.

The conversion of water peaked at XH2O = 7.1% for the cycles evaluated within the modelling framework. Reactor implementation in an FT plant require a relatively dry feed and the coupling of the two concepts seems improbable at this stage.

The biggest uncertainties tied to the results is the aforementioned lack of a true system optimization and uncertainties tied to the validity of the kinetic model at elevated pressure (5 bar).