Hydrogen production from steam–methanol reforming: thermodynamic analysis

https://doi.org/10.1016/S0360-3199(99)00013-0Get rights and content

Abstract

Thermodynamic equilibrium involved in the steam reforming of methanol is re-examined to cover the extended range of compounds suggested by literature to be involved in the reactions. The equilibrium concentrations are determined for different mixtures of these compounds at 1 atm and at different temperatures (360–573 K) and at different steam/methanol molar feed ratios (0–1.5), by the method of direct minimization of Gibbs free energy. The possibility of carbon formation in these conditions is determined by direct inclusion of carbon in the objective function of the minimization scheme. Results showed that the area of carbon formation region is surprisingly high. Carbon and methane formations are thermodynamically favoured and they reduce the quantity and quality of hydrogen produced. Dimethyl ether formation occurs at low temperatures and low steam/carbon feed ratios, while carbon monoxide occurs at high temperatures and low steam carbon ratios.

Introduction

Currently, increasing attention is being paid to the low temperature steam reforming of methanol to produce high purity hydrogen to use as fuels in fuel cells for on-board power generation for vehicles [1], [2]. The favour for methanol as a chemical carrier for hydrogen is mainly due to its ready availability, high energy density, and easy storage and transportation [2], [3], [4].

The most widely used catalysts for this reaction are copper containing catalysts since copper has been found to be highly active and selective for hydrogen [5], [6]. According to Refs [5], [6], [7], [8], [9], [10], various types of copper catalysts have been used to give slightly different components and concentrations and hence, different mechanisms to the overall reaction:CH3OH+H2O=CO2+3H2

This shows the various influences of the different types and preparation methods of heterogeneous catalysts on thermodynamic equilibrium. In this study, thermodynamics of the steam–methanol reforming system is investigated to know the equilibrium compositions within the operating range of interest. From this, ideal conditions for the reaction system to maximize hydrogen production and minimize undesirable products can be determined.

Amphlett et al. [3] examined the thermodynamics of four different models to determine the effect of carbon and methane formations on steam–methanol reforming at different temperatures, pressures and feed ratios, based on the widely accepted decomposition-shift mechanisms:CH3OH=CO+2H2CO+H2O=CO2+H2

Thermodynamic analyses prior to them were either based on the overall reaction (1) or calculated at a fixed feed ratio [3]. In all of these investigations the method of determining equilibrium concentrations was based on stoichiometry and equilibrium constants of known reactions.

Recently, Maggio et al. [11] applied such models to make comparative study of the internal steam reforming of methane, methanol and ethanol in a molten carbonate fuel cell. But carbon formation was not considered in their calculations. Most of the previous investigators (e.g. [12]) used the principle of equilibrated gas [13] to predict the carbon formation in steam–hydrocarbon reforming. Recently Vasudeva et al. [14] estimated the carbon concentration in steam–ethanol reforming by the method of Gibbs energy minimization.

In our investigation we consider four different sets of possible compounds indicated by kinetics investigators using different copper containing catalysts. The temperature range of 360–573 K and steam/methanol molar feed ratios from zero (methanol decomposition) to 1.5 (excess steam) are used in the investigation. The lower limit of the temperature range considered is 360 K since it is about the operating temperature of a typical Proton Exchange Membrane Fuel Cell (PEMFC) to which the steam reformer is to be coupled. The pressure is kept constant at 1 atm (101.32 kPa) since previous investigations [2], [3], [4] have already shown that the higher pressures are thermodynamically not favoured by steam–methanol reforming. The method of direct Gibbs energy minimization is used to determine the equilibrium concentrations. The carbon formation is also directly estimated by describing a way of incorporating carbon concentration into the objective function of the minimization scheme.

Section snippets

Thermodynamic analysis

The total Gibbs free energy of a reacting system reaches a minimum at equilibrium. The total Gibbs function for a system is given bynG=i=1NniḠi=niG0i+RTni lnf̂if0i

For gas phase reactions, f̂i=φ̂iyiP. Although the fugacity is closed to the pressure at the condition of the calculation (1 atm), we include it for a more general case.

Since the standard state is taken as the pure ideal gas state at 1 atm, f0i=1 atm, and since G0i is set equal to zero for each chemical element in its standard

Results and discussion

The results show that the equilibrium concentrations of HCOOCH3, HCHO, and HCOOH are zero for all cases. This means that even if these components are involved in the reaction mechanisms they are merely intermediates. From the Gibbs energy values results, it is found that the decomposition of methanol is thermodynamically less favoured than steam reforming for all cases. Also the methanol conversion and selectivity for hydrogen are higher in steam reforming than in methanol decomposition.

Conclusions

  • 1.

    If carbon and methane formations are not considered, the thermodynamic optimum condition for hydrogen production occurs at 1 atm pressure, 400 K and a steam/methanol feed ratio of 1.5. Under this condition the equilibrium concentration of CO is less than 1000 ppm and that of DME is less than 300 ppm, with a hydrogen yield of 2.97 moles per mole of methanol and methanol conversion of 99.7%.

  • 2.

    Although the concentrations of CO and DME can be further reduced at feed ratios greater than 1.5, H2 yield

Glossary

DLCONF Double precision version of the subroutine LCONF (linearly constrained minimization of general objective functions with finite difference gradient).

DNEQNF Double precision version of the subroutine NEQNF (nonlinear equations solver with finite difference Jacobian).

IMSL International Mathematical and Statistical Library.

LCONG Linearly constrained minimization of general objective function with analytic gradient.

TOLMIN A tolerant algorithm for linearly constrained optimization calculations.

Acknowledgements

The financial support provided by the Ministry of Science, Technology and Environment of Malaysia under the Project IRPA 02-02-02-0002 is gratefully acknowledged.

References (23)

  • S.G. Chalk et al.

    The US Department of Energy—investing in clean transport

  • Cited by (144)

    View all citing articles on Scopus
    View full text