skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: DIFFUSION OF H THROUGH PD MEMBRANES EFFECTS OF NON-IDEALITY ON DH AND ED

Abstract

H diffusion constants, D{sub H}, have been obtained from steady-state fluxes across Pd membranes with the downstream side maintained at p{sub H2} {approx} 0. Good linearity of plots of H flux versus (1/d), where d is the thickness, attests to the H permeation being bulk diffusion controlled in this temperature (423 to 523K) and p{sub H2} range ({le} 0.2 MPa). D{sub H} values have been determined at constant p{sub up} and also at constant (H/Pd)=r conditions. H fluxes through Pd membranes with three different surface treatments have been investigated (polished (un-oxidized), oxidized, and palladized) in order to determine the effects of these pretreatments. The palladized and oxidized membranes give similar D{sub H} values but the polished membranes give values about 12% lower. For diffusion in a concentration gradient D{sub H}*(c{sub H}/RT)(d{mu}{sub H}/dx) is the more proper description, where c{sub H} is the H concentration, rather than D{sub H}(dc{sub H}/dx) where D{sub H} and D{sub H}* are the concentration-dependent and independent diffusion constants. D{sub H}* can be obtained from D{sub H} using the thermodynamic factor, D{sub H}(r) = D{sub H}*({partial_derivative}lnp{sub H2}{sup 1/2}/{partial_derivative}lnr){sub T} = D{sub H}*f(r). In the commonly employed situation where there is a large difference in concentrations between themore » upstream and downstream sides of a membrane, the thermodynamic factor varies with distance through the membrane and this should be allowed for in obtaining D{sub H}*. Procedures are given and utilized for using D{sub H}(c{sub H}) to determine D{sub H}* values when there is a large concentration gradient through the membrane. Activation energies for diffusion, E{sub D}(c{sub H}), have been determined. E{sub D} is found to increase with c{sub H} which can be attributed to the thermodynamic factor. D{sub H}* values have been found to increase with H content.« less

Authors:
Publication Date:
Research Org.:
SRS
Sponsoring Org.:
USDOE
OSTI Identifier:
901095
Report Number(s):
WSRC-STI-2007-00125
Journal ID: ISSN 0925-8388; JALCEU; TRN: US200713%%62
DOE Contract Number:
DE-AC09-96SR18500
Resource Type:
Journal Article
Resource Relation:
Journal Name: J. of Alloys and Compounds or J. of Membrane Science
Country of Publication:
United States
Language:
English
Subject:
08 HYDROGEN; DIFFUSION; MEMBRANES; SURFACE TREATMENTS; THERMODYNAMICS; THICKNESS; HYDROGEN; MEMBRANE TRANSPORT; PALLADIUM

Citation Formats

Shanahan, K. DIFFUSION OF H THROUGH PD MEMBRANES EFFECTS OF NON-IDEALITY ON DH AND ED. United States: N. p., 2007. Web.
Shanahan, K. DIFFUSION OF H THROUGH PD MEMBRANES EFFECTS OF NON-IDEALITY ON DH AND ED. United States.
Shanahan, K. Fri . "DIFFUSION OF H THROUGH PD MEMBRANES EFFECTS OF NON-IDEALITY ON DH AND ED". United States. doi:. https://www.osti.gov/servlets/purl/901095.
@article{osti_901095,
title = {DIFFUSION OF H THROUGH PD MEMBRANES EFFECTS OF NON-IDEALITY ON DH AND ED},
author = {Shanahan, K},
abstractNote = {H diffusion constants, D{sub H}, have been obtained from steady-state fluxes across Pd membranes with the downstream side maintained at p{sub H2} {approx} 0. Good linearity of plots of H flux versus (1/d), where d is the thickness, attests to the H permeation being bulk diffusion controlled in this temperature (423 to 523K) and p{sub H2} range ({le} 0.2 MPa). D{sub H} values have been determined at constant p{sub up} and also at constant (H/Pd)=r conditions. H fluxes through Pd membranes with three different surface treatments have been investigated (polished (un-oxidized), oxidized, and palladized) in order to determine the effects of these pretreatments. The palladized and oxidized membranes give similar D{sub H} values but the polished membranes give values about 12% lower. For diffusion in a concentration gradient D{sub H}*(c{sub H}/RT)(d{mu}{sub H}/dx) is the more proper description, where c{sub H} is the H concentration, rather than D{sub H}(dc{sub H}/dx) where D{sub H} and D{sub H}* are the concentration-dependent and independent diffusion constants. D{sub H}* can be obtained from D{sub H} using the thermodynamic factor, D{sub H}(r) = D{sub H}*({partial_derivative}lnp{sub H2}{sup 1/2}/{partial_derivative}lnr){sub T} = D{sub H}*f(r). In the commonly employed situation where there is a large difference in concentrations between the upstream and downstream sides of a membrane, the thermodynamic factor varies with distance through the membrane and this should be allowed for in obtaining D{sub H}*. Procedures are given and utilized for using D{sub H}(c{sub H}) to determine D{sub H}* values when there is a large concentration gradient through the membrane. Activation energies for diffusion, E{sub D}(c{sub H}), have been determined. E{sub D} is found to increase with c{sub H} which can be attributed to the thermodynamic factor. D{sub H}* values have been found to increase with H content.},
doi = {},
journal = {J. of Alloys and Compounds or J. of Membrane Science},
number = ,
volume = ,
place = {United States},
year = {Fri Mar 09 00:00:00 EST 2007},
month = {Fri Mar 09 00:00:00 EST 2007}
}
  • Under the commonly employed experimental conditions of a significant upstream concentration of H and c{sub H} {approx} 0 downstream, expressions are given for obtaining the concentration-independent D*{sub H} from the concentration dependent D{sub H} employing the known non-ideality. A procedure is given for determining the concentration profile for a given upstream concentration for an alloy where the non-ideality is known as a function of H concentration. For the Pd{sub 0.81}Ag{sub 0.19} alloy (423 K) the nonideality, f(r)<1 decreases the flux but for alloys where the non-ideality is in the opposite direction, f(r)>1, the flux will be greater which would bemore » an advantage for the experimental purification of H{sub 2}.« less
  • In 1998 Kirchheim et al remarked that ''to their knowledge, experimental results on the diffusion of hydrogen through multi-layers have not yet been reported'' [1]. Their research dealt with diffusion through ultra-thin multi-layers of Nb/Pd which they followed electrochemically using a time-lag method. Their results were somewhat uncertain in that no final conclusion about any effect of the internal interfaces could be reached. Very recently Yamakawa et al [2] investigated Pd/Fe and Pd/Ni multilayers at 378-625 K and found no strong influence of the interfaces, however, there was grain boundary diffusion for the Pd/Ni layers and retardation, possibly due tomore » dislocation trapping. Takano et al [3] studied H diffusion through thin layers of Pd, Ni, or Cu deposited electrochemically on Fe and concluded that complications are introduced by thin films perhaps H trapping at vacancies as the thickness of the layer decreases to very small values. Holleck [4] measured H diffusion through mm thick Pd{sub 0.75}Ag{sub 0.25}/Ta/Pd{sub 0.75}Ag{sub 0.25} layers in the gas phase from 540-873 K and determined D{sub H,Ta} from the overall diffusion constant and the known D{sub H,alloy}; he concluded that the interface did not play a significant role at these temperatures. As in Holleck's, the present research does not concern very thin multi-layers but macroscopic ones, in this case, prepared by internal oxidation. The measurements will be carried out at a lower temperature than Holleck's and with a variety of layer thicknesses. Most previous investigations have employed electrochemical time-lag methods [1,3,5,6] at ambient temperature and have assumed that the solubilities at the interfaces are in the ideal range. To our knowledge, an investigation of diffusion through layers prepared by internal oxidation has not been carried out. Partial internal oxidation of Pd-M alloys leads to outer layers of Pd containing M oxide precipitates while the inner layer remains unoxidized Pd-M alloy. The alloy employed in this research is Pd{sub 0.96}Al{sub 0.04} which was chosen because the authors have had extensive experience with its internal oxidation [7,8] and it has been shown that internally oxidized Pd-Al alloys are more resistant to poisoning by CO than pure Pd [9]. The specific Al content was chosen in order for it to be large enough for the alloy to have a significantly lower permeability than Pd but small enough to internally oxidize. The diffusion results may give some information about internal oxidation.« less