Analysis of oxygen permeation through dense ceramic membranes with chemical reactions of finite rate

https://doi.org/10.1016/j.ces.2008.09.022Get rights and content

Abstract

The oxygen permeation through oxygen ionic or mixed-conducting ceramic membranes under reaction conditions was examined with a model taking into account of different electrical transport mechanisms (p-type and n-type transports) and finite reaction rate. It was demonstrated that with a reaction consuming oxygen in one side of the membrane, the oxygen partial pressure in the reaction side decreases and the oxygen permeation flux increases with the increase in the reaction rate for both the p-type and the n-type transport dominated mechanism. The increase in reaction rate causes a transition of the transport mechanism from p-type to n-type. This transition leads to an increase in the permeation flux by up to 30 times. This effect offers one explanation for the large discrepancies in published permeation data for membrane reactors of partial oxidation reaction employing an oxygen permeable dense ceramic membrane. For a membrane with a specific transport mechanism, the increase in the reactant partial pressure causes an increase in the reaction rate and oxygen permeation flux. However, the increase in the inlet inert gas amount has a complicated effect on the oxygen permeation flux because it lowers both oxygen partial pressure and the reaction rate at the same time.

Introduction

Dense fast ion-conducting ceramic membranes are of considerable interest for academia and industry (Kilner, 2000; Lin, 2001). Those are applied extensively as a tunable oxygen separation media in electrocatalytic and membrane reactors either for power generation or conversion of light hydrocarbons to value-added products. Particularly, the mixed ionic and electronic conducting perovskite structured lanthanum membranes (Balachandran et al., 1995; ten Elshof et al., 1995a; Tsai et al., 1997; Xu and Thomson, 1998; Zeng et al., 1998), and fluorite (Fagg et al., 2007) or fluorite-like (Horovistiz et al., 2007) structured ceramic membranes have been investigated most extensively in the past decade. However, inconsistent oxygen permeation fluxes, with difference as large as one order of magnitude, were reported frequently even for a nominally identical membrane material. Teraoka et al. (1985) reported an oxygen permeation flux as high as 2.31×10−6 mol cm−2 s−1 at 850 °C for a SrCo0.8Fe0.2O3−δ membrane; while the data for the same membrane with a same thickness measured under similar experimental conditions from Kruidhof et al. (1993) and Qiu et al. (1995) were 1.8×10–7 and 6.3×10−7 mol cm−2 s−1, respectively. Conflicting results were also reported concerning the effect of A-site substitution in the perovskite by different cations. Teraoka et al. (1988) and Tsai et al. (1998) showed that the oxygen permeation fluxes of A-site substituted LaCo1−yFeyO3−δ membrane had an order Ba>Ca>Sr, however, Stevenson et al. (1996) and Li et al. (1999) reported a different order Sr>Ba>Ca.

Several research groups have noticed and tried to explain the discrepancies. Differences in experimental details, such as the gas atmosphere which modifies the membrane surface (Yi et al., 2005; Arnold et al., 2007), the sealing material (Qiu et al., 1995) and the difference in preparation procedure (Kharton et al., 2001; Qi et al., 2000; Wu et al., 2006; Zhang et al., 1999) have been suggested to account for such discrepancies. Yi et al. (2005) found that the co-presence of CO2 and H2O in air had profound effects on the phase composition, microstructure, and oxygen permeability of a Sr0.95Co0.8Fe0.2O3−δ membrane, and attributed these effects to the formation of bicarbonate on the membrane surface. Wu et al. (2006) examined the effect of pH during synthesis via a citrate process on the properties of the resulting perovskite membrane and reported that the crystallinity and the oxygen permeability of the La0.6Sr0.4Co0.4Fe0.6O3−δ membrane can be tailored. Qi et al. (2000) investigated the electrical conductivity and the oxygen permeability of a La0.8Sr0.2Co0.6Fe0.4O3−δ membrane prepared from powders synthesized by four different techniques, i.e. the citrate, solid state, spray-pyrolysis and coprecipitation methods. They found that a substantial deviation from the desired stoichiometry existed for the membranes prepared by different methods and proposed that this deviation was the major reason for the discrepancies in the reported oxygen permeability data for the membranes with similar composition.

Akin and Lin (2004) recently analyzed the effect of downstream conditions on oxygen permeation flux through ionic and mixed-conducting ceramic membranes by a simple mathematical model for special cases with extremely slow and fast reaction rate. They reported that the downstream gas reactivity with oxygen and flow rate and pattern can have a strong effect on oxygen permeation flux through these membranes. This work not only examined the effects of the operating conditions on oxygen permeation flux but also provided the direction to study oxygen permeation through this group of membranes under more realistic reaction conditions with a finite reaction rate. Such a study is very important to explain the discrepancy in oxygen permeation flux data and development of ionic or mixed-conducting ceramic membrane reactors for partial oxidative reactions. The present paper reports the study of oxygen permeation through this group of membranes with a reaction of finite rate in the membrane downstream by a much improved mathematical model considering oxygen permeation, reactor configuration and reaction kinetics.

Section snippets

Model development

In membrane reactors, an ionic or mixed-conducting ceramic membrane separates an oxygen containing upstream (typically air) and a downstream containing a reactant, such as methane, with or without an inert carrier gas, as schematically illustrated in Fig. 1.The reactant in the downstream reacts with oxygen permeating through the membrane from the upstream. The driving force for oxygen permeation is the Gibbs energy for the reaction but is often stated in the literature as the difference between

Results and discussion

The reaction side oxygen partial pressure is dependent on the constants chosen in Eq. (1). Here, we use the parameters for Bi1.5Y0.3Sm0.2O3−δ (BYS) to represent a p-type membrane and those for La0.6Sr0.4Co0.2Fe0.8O3−δ (LSCF) to represent an n-type membrane, as listed in Table 1. It should be noted that the use of term “n- or p-membrane” is to indicate membrane materials whose oxygen permeation flux equations can be described by either Eq. (2) or Eq. (3). It does not necessarily mean the

Conclusions

An oxygen permeation membrane reactor model taking into account of the different transport mechanisms, i.e. with p-type and n-type flux equations was developed to analyze the oxygen permeation through an oxygen ionic or mixed-conducting ceramic membrane under reaction conditions. The simulation results show that the oxygen partial pressure in the reaction side decreases with the increase in the reaction rate for both the p-type and n-type membranes. As a consequence, for both the two kinds of

Notation

Aeffective oxygen permeation area, cm2
FFaraday constant, C mol−1
jO21,jO22dimensionless oxygen permeation parameter, jO21=Ak1PO21/n1/Qfeed0, jO22=Ak2PO21/n2/Qfeed0
JO2oxygen permeation rate, mol cm−2 min−1
k1, n1,constant
k2, n2,
kCOkinetic parameter, mol min−1 ml−1 atm−1.5
kCOdimensionless kinetic parameter, kCO=kCOVPt1.5/Qfeed0
Lmembrane thickness, cm
poxygen partial pressure ratio between the reaction side and air side, p=PO2/PO2
Ppressure, atm
PO2oxygen partial pressure at the air side, atm
PO2

Acknowledgments

This work has been supported by the Natural Science Foundation of China under Contract numbers 20425619 and 20736007. The work has been also supported by the Program of Introducing Talents to the University Disciplines under File number B06006, and the Program for Changjiang Scholars and Innovative Research Teams in Universities under File number IRT 0641.

References (34)

  • W.T. Stephens et al.

    Influence of gas flow rate on oxygen flux measurements for dense oxygen conducting ceramic membranes

    Solid State Ionics

    (2000)
  • J.E. ten Elshof et al.

    Oxidative coupling of methane in a mixed conducting perovskite membrane reactor

    Applied Catalysis A

    (1995)
  • J.E. ten Elshof et al.

    Oxygen transport through La1−xSrxFeO3−δ membranes I. Permeation in air/He gradients

    Solid State Ionics

    (1995)
  • B.A. van Hassel et al.

    Oxygen permeation flux through La1−ySryFeO3 limited by carbon monoxide oxidation rate

    Applied Catalysis A

    (1994)
  • Y. Zeng et al.

    Perovskite type ceramic membrane: synthesis, oxygen permeation and membrane reactor performance for oxidative coupling of methane

    Journal of Membrane Science

    (1998)
  • H.J.M. Bouwmeester et al.

    Dense ceramic membranes for oxygen separation

  • Fagg, D.P., Shaula, A.L., Kharton, V.V., Frade, J.R., 2007. High oxygen permeability in fluorite-type Ce0.8Pr0.2O2-δ...
  • Cited by (40)

    • Modelling and sequential simulation of multi-tubular metallic membrane and techno-economics of a hydrogen production process employing thin-layer membrane reactor

      2016, International Journal of Hydrogen Energy
      Citation Excerpt :

      However, coupling a selective membrane separation process with chemical reactions and mass transfer on both sides of reactor in MR, presents interesting challenges [26–29]. During past decades there has been numerous theoretical attempts to model MRs based on different assumptions and for different membrane configurations [9,28,30–33]. But, these models have been under restriction of considered assumptions and operating parameters and there is no developed single validated SMR-MRs model [34].

    • Autothermal reforming of ethanol in dense oxygen permeation membrane reactor

      2016, Catalysis Today
      Citation Excerpt :

      The reactants in the downstream react with oxygen permeating through the membrane from the upstream. The driving force for oxygen permeation is the Gibbs energy for the reaction but is often stated in the literature as the difference between the oxygen partial pressure in the upstream and downstream [22,23]. In our EOSR reaction in DMR experiment, the downstream with reactants reacting with oxygen lower the oxygen partial pressure, thus enhancing the driving force for oxygen permeation and hence oxygen permeation flux by a factor of 8–10 times as compared to that under air/Ar gradient case.

    View all citing articles on Scopus
    View full text