i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 3 4 7 7 e3 4 9 0
Available online at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/he
Characterization of PdeCu membranes fabricated by
surfactant induced electroless plating (SIEP) for hydrogen
separation
M.S. Islam, M.M. Rahman, S. Ilias*
Department of Chemical, Biological and Bioengineering, North Carolina A&T State University, Greensboro, NC 27411, USA
article info
abstract
Article history:
PdeCu composite membranes on microporous stainless steel (MPSS) substrate were
Received 26 August 2011
fabricated using surfactant induced electroless plating (SIEP). In the SIEP method, dodecyl
Received in revised form
trimethyl ammonium bromide (DTAB), a cationic surfactant, was used in Pd- and Cu-baths
7 November 2011
for the sequential deposition of metals on MPSS substrates. The SIEP PdeCu membrane
Accepted 9 November 2011
performance was compared with membranes fabricated by conventional electroless
Available online 6 December 2011
plating (CEP). The pre- and post-annealing characterizations of these membranes were
carried out by SEM, XRD, EDX and AFM studies. The SEM images showed a significant
Keywords:
improvement of the membrane surface morphology, in terms of metal grain structures and
SIEP
grain agglomeration compared to the CEP membranes. The SEM images and helium gas-
CEP
tightness studies indicated that dense and thinner films of PdeCu can be produced with
CMC
shorter deposition time using SIEP method. From XRD, cross-sectional SEM and EDS
Permeability
studies, alloying of PdeCu was confirmed at an annealing temperature of 773 K under
H2-selectivity
hydrogen environment. These membranes were also studied for H2 perm-selectivity as
Surfactant DTAB
a function of temperature and feed pressure. SIEP membranes had significantly higher H2
perm-selectivity compared to CEP membranes. Under thermal cycling (573 K e 873 K e
573 K), the SIEP PdeCu membrane was stable and retained hydrogen permeation characteristics for over three months of operation.
Copyright ª 2011, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights
reserved.
1.
Introduction
The issue of climate change has prompted scientific and
policy efforts to limit global carbon releases and has raised
interest in emission-free energy productions [1e5]. Hydrogen
continues to be an important industrial feedstock, and is
becoming a fuel of choice for green energy [3,6e11]. In light of
the rapid development of hydrogen-based energy and the
high-technology industry, hydrogen separation technologies
are becoming increasingly important. Currently membrane
separation is one of the most important technologies for high
purity hydrogen [6,12e14]. Specially, for high temperature
applications, Pd-based membranes have been the focus of
many studies because of their high selectivity of hydrogen gas
from other gases due to the high solubility and mobility of
hydrogen in the Pd lattice [15e18]. However, hydrogen
embrittlement of the Pd-layer is a problem with pure Pd
membranes. Consequently, it cannot be used for H2-separation below 573 K due to the lattice expansion caused by dissolved hydrogen [3,19e21]. Further, cracking of the Pd
* Corresponding author. Tel.: þ1 336 334 7564.
E-mail address: [email protected] (S. Ilias).
0360-3199/$ e see front matter Copyright ª 2011, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijhydene.2011.11.024
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 3 4 7 7 e3 4 9 0
Table 1 e Chemical composition of cleaning solution.
Component
Supplier
Composition
Na3PO4$12H2O
(ACS reagent grade, 99.4%)
Na2CO3
(ACS reagent grade, 99.5%)
NaOH
(ACS reagent grade, 97%)
Industrial detergent
(Liqui-NoxR)
Alfa Aesar
45 g/L
Alfa Aesar
65 g/L
Alfa Aesar
45 g/L
Alconox
5 mL/L
membrane under thermal stress due to the aeb Pd hydride
phase transition, as well as poisoning or fouling due to the
presence of sulfur or unsaturated carbon compounds in the
operating stream, severely limits its use [19e21]. To alleviate
these shortcomings, Pd-based membranes have been made
from alloying with metals, such as silver, gold, nickel and
copper [19,20,22]. By alloying Pd with selected metals, significant H2-permeability enhancement can be achieved. For
example, the Pd77Ag23 and Pd60Cu40 (wt.% composition) alloys
show 73.4% and 6.3% increase in permeability respectively,
over the pure Pd system [19,22e29]. Alloy materials, such as
copper or gold, are more resistant to sulfur compounds
[19,21,22]. Specially, the PdeCu alloy film has enhanced
thermal resistance without suffering any discernible physical
changes [3,5,20,30]. The H2-permeability of the Pd-Cu
membrane passes through a maximum around 40 wt.% of
Cu, in particular offering a cost reduction benefit
[3,19,21,22,28]. In addition, Pd60Cu40 can withstand repeated
temperature cycling with less distortion than pure Pd at or
below room temperature [3,9,31].
Pd and PdeCu membranes can be fabricated by various
methods including chemical vapor deposition, sputtering
coating, electrochemical plating and electroless plating (EP)
[5,8]. Of all these fabrication methods, EP is most widely used
because of the advantages it has over other processes [32]. Roa
et al. fabricated H2-selective Pd80Cu20 and Pd78Cu22
membranes on a microporous ceramic support using the EP
method with metal film thickness in the range of 10e13.5 mm.
Table 2 e Chemical composition of sensitization and
activation solutions.
Component
Supplier
SnCl2$2H2O
SigmaeAldrich
(ACS Reagent
grade, 98%)
PdCl2
Alfa Aesar
(ACS reagent
grade, 99.9%)
HCl
SigmaeAldrich
(ACS reagent
grade, 37%)
Operating conditions
Temperature
Time
pH
Sensitization
solution
Activation
solution
1 g/L
e
e
0.1 g/L
1 mL/L
1 mL/L
Table 3 e Chemical composition of Pd-bath and Cu-bath
solutions.
Name of Chemicals
Pd (NH3)4Cl2$H2O
(99.99%)
CuSO4$5H2O
(ACS grade,
98.0e102.0%)
Na2EDTA (99%)
NH4OH
(ACS grade, 29.17%)
NaOH
(ACS grade, 97%)
N2H4 (1.0 M)
DTAB (w99%)
Formaldehyde
(ACS, 37% in aq. soln)
2, 20 - Bipyridine
(>99%)
Operating conditions
Time
Temperature
pH
Supplier
Pd-bath
Cu-bath
SigmaeAldrich
4 gm/L
e
Alfa Aesar
e
1.22 gm/L
Acros Organics
Fisher Scientific
40.1 gm/L
198 mL/L
10.05 gm/L
e
Alfa Aesar
e
10 gm/L
SigmaeAldrich
SigmaeAldrich
Alfa Aesar
5.6 mM
4 CMC
e
e
4 CMC
280 mL
Alfa Aesar
e
2.5 g/L
1h
333 K
10e11
20 min
333 K
12e13
The PdeCu composite membranes were tested for H2-separation at temperatures between 623 K and 973 K [3,28]. Using
the EP method, Pd75Cu25 membranes on ceramic support with
a film thickness of about 2e4 mm were fabricated by Kulprathipanja et al. for the separation of hydrogen from a water gas
shift reaction at a temperature range of 623e723 K [33]. Yang
600
S3 Pd-Cu Membrane (SIEP)
S2 Pd-Cu Membrane (SIEP)
C1 Pd-Cu Membrane (CEP)
C2 Pd-Cu Membrane (CEP)
500
Helium flow rate (m3/m2-s)
3478
400
300
200
100
0
0
5
10
15
20
25
Film Thickness (µ m)
293 K
4e6 min
4e5
293 K
4e6 min
4e5
Fig. 1 e Helium gas-tightness of PdeCu membranes as
a function film thickness fabricated by CEP and SIEP
methods.
30
3479
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 3 4 7 7 e3 4 9 0
Table 4 e Permeation characteristics of Pd and PdeCu membranes fabricated by SIEP and CEP methods.
Membrane
sample
Deposition
time (hr)
Film thickness
(mm)
Gravimetric
analysis
S1 PdeCu
S2 PdeCu
S3 PdeCu
S4 PdeCu
C1 PdeCua
C2 PdeCua
Pd 4CMCb
Pd 0CMCb
26.8
33.8
31.4
27.8
34.3
35.7
10
28
15.6
16.73
14.5
14.7
19.18
20.17
7.68
28.5
SEM
analysis
e
e
13.34
13.61
e
e
7.68
27.5
Film
composition
(% of Cu)
Sieverts’
law power
index (n)
e
35.8
44.69
25.8
36.7c
33.4c
e
e
e
0.8
0.79
e
0.92
0.91
0.61
e
H2 flux at 723 K
(mol/m2-s)
Selectivity at 723 K
(H2 flux/N2 flux)
20 psi
100 psi
20 psi
100 psi
e
0.145
0.156
e
0.094
0.097
0.28
e
e
0.54
0.55
e
0.43
0.44
0.81
e
e
123
110
e
19
17
250
e
e
49
43
e
7.42
8
e
e
a Membranes fabricated by CEP.
b Data from previous work [39].
et al. investigated the effect of Cu content in the PdeCu
system on hydrogen permeability. The Pd100xCux (x ¼ 0 to
60 wt.% Cu) thin-films of about 200 nm were deposited on V15Ni support by the DC co-sputtering method. They claimed
that the Pd60Cu40/V-15Ni membrane displayed higher
hydrogen permeability than the Pd60Cu40 composite
membrane on porous support [6,7,31]. To control the PdeCu
alloy composition, Hoang et al. used the dual sputtering
method to fabricate a PdeCu membrane on silicon support
with a film thickness of 750 nm [8]. Gao et al. introduced a thin
layer of ZrO2 as an intermetallic diffusion barrier in a Pd46Cu54
membrane fabricated by the EP technique on microporous
stainless steel (MPSS) discs with a film thickness of 10 mm. It is
claimed that the diffusion barrier improved the membrane
stability in practical applications [5]. Recently, Pomerantz and
Ma reported the fabrication of a Pd/Cu/Pd tri-layer membrane
of 19 mm thick on PSS tubular support by the EP method by
rotating the support in the plating bath for improved deposition rates [4].
Chen et al. studied the role of surfactant in a nickelphosphorus (Ni-P) EP bath and observed that a suitable
surfactant can increase the deposition rate by up to 25% in
brass substrates [34]. This study established a quantitative
and qualitative relationship between the effect of surfactants
and the surface properties of the Ni-P plating layer, as well as
corrosion resistance of the resulting Ni-P deposits.
The long-term performance of the Pd-composite membrane
fabricated by the EP method is greatly affected by the crystallite
Fig. 2 e SEM images of PdeCu membrane film top surface fabricated by SIEP method.
3480
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 3 4 7 7 e3 4 9 0
35
Cu
distribution and microstructural characteristics of the film [35].
The oxidationereduction reactions between the Pd-complex
and hydrazine in the EP result in metallic deposition of Pd in
a solid surface by releasing NH3 and N2 gas bubbles. The gas
bubbles adhering to the substrate surface and in the pores
hinder uniform Pd-film deposition. To address this problem,
we investigated the role of surfactant in a Pd EP bath and
demonstrated that surfactant induced EP (SIEP) can effectively
tailor and control the Pd deposition rate, grain size and grain
agglomeration on an MPSS substrate [32,36]. We extended the
SIEP method to a Cu-bath to fabricate a Pd-Cu membrane by
sequential deposition of Pd and Cu on an MPSS support. In this
paper we report the fabrication and characterization of the
PdeCu composite membrane by the SIEP method.
4CMC DTAB
30
Relative abundance (%)
25
20
15
Pd
2.
10
5
0
0
1
2
3
4
Particle Size (µm)
Fig. 3 e Cu and Pd grain size distribution in PdeCu
membrane fabricated by SIEP method with 4 3 CMC of
DTAB surfactant.
5
Materials and methods
EP is a controlled autocatalytic deposition of a continuous film
on a catalytic interface by the reaction of a metal salt and
a chemical reducing agent [37]. Typical bath chemistry for the
EP of Pd or Cu consists of a metal ion source, a complexing
agent, and a reducing agent. In this work, we used the Pd- and
Cu-EP baths as follows:
Pd-EP Bath Reactions
o
2PdðNH3 Þ2þ
4 þ 4e /2Pd þ 8NH3
N2 H4 þ 4OH /N2 þ 4H2 O þ 4e
Overall Reaction:
o
2PdðNH3 Þ4 Cl2 þ N2 H4 þ 4NH4 OH/2Pd þ N2 þ 8NH3
þ4NH4 Cl þ 4H2 O
Fig. 4 e SEM images of top surface of PdeCu membranes fabricated by CEP and SIEP methods at pre- and post-HT (heat
treatment) conditions.
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 3 4 7 7 e3 4 9 0
Cu-EP Bath Reactions
o
CuðNH3 Þ2þ
6 þ 2e /Cu þ 6NH3
2HCHO þ 4OH /2HCOO þ 2H2 O þ H2 þ 2e
Overall Reaction:
CuðNH3 Þ6 SO4 þ 2HCHO þ 4NaOH/Cuo þ 6NH3 þ 2HCOONa
þNa2 SO4 þ 2H2 O þ H2
The membrane support, MPSS substrates (316 L SS discs, 1
inch diameter, 0.062 inch thickness with an average pore size
of 0.2 mm) were obtained from Mott Metallurgical Corporation
(Farmington, CT). Prior to the sensitization and activation
processes, the substrate surface was thoroughly cleaned in
three steps. The substrate was cleaned by dipping it into an
alkaline cleaning solution (composition is given in Table 1) for
40e60 min at 333 K in an ultrasonic bath followed by a thorough cleaning with deionized water until the pH of the
substrate surface reached level 7. Finally, the substrate was
dipped into an isopropanol solution (Fisher Scientific) for
10 min, and dried for 2 h at 393 K.
The sensitization and activation solutions were prepared
using reagent grade SnCl2 and PdCl2. The substrate surface
was sensitized and activated sequentially by dipping it into
SnCl2 and PdCl2 solutions, respectively. The substrate was
3481
rinsed with deionized water between the sensitization and
activation steps. The composition and conditions used in
sensitization and activation are given in Table 2.
Pd and Cu were deposited sequentially on the activated
MPSS surface by the SIEP process using a dodecyl trimethyl
ammonium bromide (DTAB) surfactant. In Table 3, composition and operating conditions of the Pd- and Cu-EP baths used
in this study are given. The DTAB concentration is expressed
in critical micelle concentration (CMC). The 4 CMC
concentration of the DTAB was used for both Pd- and Cu-baths
in SIEP process. A similar DTAB concentration was also used
in fabricating Pd and PdeAg alloy membranes in our previous
works [32,38,39]. EP without surfactant is referred to here as
conventional electroless plating (CEP).
The microstructure of the fabricated PdeCu membranes
was analyzed using scanning electron microscopy (SEM),
energy dispersive spectroscopy (EDS), X-ray diffraction (XRD)
and atomic force microscopy (AFM) techniques. The grain
sizes were determined using point-to-point measurements
from representative SEM images. The statistical distributions
of deposited metal grains were estimated considering
a minimum 500 number of grains in a constant cross-section
area. The membranes were cut into pieces to analyze the
cross-section across the thickness of the PdeCu alloy film
Fig. 5 e AFM images of solid PdeCu surface aggregation onto typical MPSS (SIEP and CEP).
3482
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 3 4 7 7 e3 4 9 0
8000
Element
Pd
Cu
Wt. %
55.31
44.69
Std. dev. Atomic %
1.42
42.5
1.32
57.5
6000
Counts
Pd
4000
Cu
Pd
2000
Cu
Pd
Pd
0
0
2
4
6
8
10
X-Ray Energy (KeV)
Fig. 6 e Typical EDS spectrum of PdeCu membrane shows
the presence of polycrystalline deposition of Pd and Cu
particles.
samples fabricated by CEP (C1 and C2) methods. The results
show that, in order to obtain a gas-tight PdeCu membrane, we
need about 15 mm thick metal film by the SIEP method. In the
CEP method, this thickness is over 19 mm. Thus, we observed
that using DTAB can reduce the membrane film thickness
with improved gas-tightness.
Electroless plating time, film thickness, metal composition,
H2 flux and selectivity data of PdeCu membranes fabricated by
the SIEP and CEP methods are summarized in Table 4. We also
include Pd membrane data from our previous work for
comparison in Table 4 [39]. The film thickness of the PdeCu
membrane was calculated using a gravimetric method and
SEM analysis for SIEP membranes. For CEP membranes, only
a gravimetric method was used to calculate the film thickness
of the PdeCu membrane. From Table 4, it is evident that PdeCu
membranes fabricated by the SIEP process are thinner
compared to PdeCu membranes fabricated by the CEP process.
The average thickness of the PdeCu film fabricated by the SIEP
process is 15.38 mm, whereas, that of PdeCu film fabricated by
the CEP process is 19.68 mm. In our previous study, we found
that the film thickness of the pure Pd membrane fabricated by
the SIEP process was 7.68 mm and the film thickness of the
PdeAg membrane fabricated by the SIEP process was 12.63 mm
[38,39]. It should be noted that PdeCu films are considerably
thicker than the Pd-film due to Cu-plating bath kinetics. The
results presented in Table 4 indicate that thinner PdeCu
membranes can be fabricated in shorter time using the SIEP
method compared to the CEP method.
3.2.
after the permeability studies. The cross-sections were
studied in terms of metal diffusivity from PdeCu film to
substrate, as well as from substrate to film using EDS mapping
and line scanning at multiple location across each crosssection. The PdeCu membranes were tested in a custombuilt diffusion cell in our permeability measurement set-up
for helium gas-tightness and H2-perm-selectivity [40].
3.
Results and discussion
3.1.
Helium gas-tightness and film thickness analysis of
PdeCu membranes
PdeCu membranes on MPSS substrate were fabricated by SIEP
and CEP methods. The microstructure of the Pd-alloy
membranes depends on a number of factors, such as
substrate surface roughness, pore dimension, fabrication
technique, bath composition and operating conditions. In this
work, all these parameters (bath composition, and operating
parameters) were kept constant in order to understand the
role of surfactant on the microstructure of the PdeCu films.
The membranes were tested for helium gas-tightness in
the permeability measurement set-up at a transmembrane
pressure drop of 20 psi at room temperature (w298 K). Helium
was chosen since it is a smaller molecule (2.6 Å) compared to
hydrogen (2.89 Å) for gas-tightness. In Fig. 1, helium gastightness results are presented for two samples of
membranes fabricated by SIEP (S2 and S3), along with two
Microstructure analysis of PdeCu membranes
The microstructure of the PdeCu film was studied using SEM
images at 5 K, 10 K, 20 K and 50 K magnifications. For example,
in Fig. 2 SEM images of a PdeCu membrane fabricated by SIEP
reveal two different particles with differing sizes. The larger
and smaller particles are mostly identified as Pd and Cu metal
clusters, respectively. The particle size distribution of this
PdeCu membrane as determined from SEM image analysis is
shown in Fig. 3. The particle size distribution of Pd and Cu
shows two sharp peaks with average particle size of 1.01 and
0.65 mm, respectively.
The PdeCu membranes fabricated by CEP and SIEP
methods were examined by SEM after annealing (heat
treatment-HT). In Fig. 4, the SEM images of PdeCu membranes
fabricated by CEP (aed) and SIEP (eef) methods are shown.
SEM images in Fig. 4(aeb)(ced) and (aed) show the PdeCu
membrane (CEP) surface morphology at pre-HT and post-HT
conditions, respectively. Pre- and post-HT SEM images of the
PdeCu (SIEP) membrane are shown in Fig. 4eef. If we compare
SEM images of the CEP membrane (a and b) at pre-HT condition with that of the SIEP membrane Fig. 4e, we observe that
the SIEP membrane has superior surface grain structure with
uniformity. Upon annealing (post-HT), we observe smooth
grain agglomeration with continuity as seen in Fig. 4f.
However, the CEP membrane show poor agglomeration with
a needle-like structure popping out of the surface as given in
Fig. 4d. However, the grain agglomeration and size of particles
are much bigger in size and shape compared to those of
PdeCu membranes fabricated by SIEP process, which is clear
in Fig. 4e and f. Thus, we observe that by introducing suitable
3483
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 3 4 7 7 e3 4 9 0
c
8000
Pd-Cu alloy{111}
Pd-Cu after 18 hrs of annealing
Intensity [a.u.]
6000
4000
Pd-Cu alloy{200}
Pd{311}
Pd-Cu alloy{220}
2000
Cu{311}
0
b
3000
Pd-Cu alloy{111}
Intensity [a.u.]
2500
Pd-Cu After 10 hrs of annealing
2000
1500
Pd-Cu alloy{200}
Pd{311}
1000
Pd-Cu alloy{220}
Cu{311}
500
0
a
1800
1600
Intensity [a.u.]
Pd-Cu Membrane before annealing
Pd{111}
1400
1200
Cu{111}
1000
800
Pd{200}
600
400
Cu{311}
Cu{220}
Cu{200}
200
Pd{311}
Pd{220}
0
40
50
60
70
80
90
2 Theta
Fig. 7 e Effect of heat treatment on XRD pattern of PdeCu membrane fabricated by SIEP method.
surfactant in the EP, grain size and structure can be tailored
and controlled to fabricate a better PdeCu film.
The surface topography of PdeCu membranes fabricated
by SIEP and CEP methods was examined by AFM analysis and
is presented in Fig. 5. The Pd-Cu membrane fabricated by SIEP
show cage-like structures (Fig. 5a) and the grains are diffused
to one another due to the presence of both Pd and Cu atoms
with average surface roughness of about 1.122 mm. For the CEP
Pd-Cu membrane (Fig. 5b), AFM analysis shows non-uniform
grain agglomeration with surface roughness of about
2.275 mm which is significantly higher than that of the SIEP
membrane. This notable improvement in the Pd-Cu film
structure is attributed to the hydrophobic DTAB used in the
SIEP method. Surfactant DTAB played an important role in
removing evolved gas bubbles from the surface in EP bath
reactions. Tail groups of DTAB appear to be aligned around the
gaseliquid interface and form various spherical or cylindrical
cage-like structures. It also tends to form various meta-stable
structures (spherical, cylindrical or circular) at the solide
liquid interface that inherently help finer grain formation and
subsequent coarsening of the deposited film. As seen in the
AFM images (Fig. 5a), DTAB forms cylindrically long, repetitive
chains throughout the surface. This observation is in agreement with other published work [32,41]. DTAB may also take
part in the reaction kinetics as it contains active bromide ions
(Br) in the head group which is a strong oxidizing agent. This
head group may participate in the reduction process of the
complex salt and favorably take part in Pd and Cu grain
3484
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 3 4 7 7 e3 4 9 0
Table 5 e Comparison of high angle XRD reflection peaks of Pd and PdeCu film fabricated by SIEP method.
Bravais lattice
2-theta
d-spacing
Lattice parameter, a
Lattice structure
111
200
220
311
111
200
220
311
Pd-film
PdeCu-film
Pd (Pre-HT)
Pd (Post-HT)
Pd (Pre-HT)
Cu (Pre-HT)
PdeCu (Post-HT)
40.214
46.776
68.303
82.338
2.2407
1.9405
1.3721
1.1701
3.881
f.c.c.
40.109
46.652
68.107
82.086
2.2463
1.9454
1.3756
1.1731
3.8908
f.c.c.
40.145
46.695
68.176
82.173
2.24439
1.94370
1.37440
1.17210
3.8874
f.c.c.
43.316
50.448
74.125
89.936
2.0871
1.8075
1.2781
1.0899
3.6150
f.c.c.
40.366
46.956
68.585
82.702
2.2326
1.9335
1.3671
1.1659
3.867
f.c.c.
formation and subsequent grain coarsening. These dual roles
of DTAB appear to be very effective in thin film formation.
The typical EDS pattern of PdeCu film fabricated by SIEP is
presented in Fig. 6. The EDS pattern shows two distinct peaks
for Pd and Cu. The elemental compositions of Pd and Cu were
found to be 55.31% and 44.69% by weight, respectively.
Absence of other metal peaks in EDS spectra in Fig. 6 indicates
that use of DTAB in the EP baths did not introduce any kind of
impurities in membrane film.
Typical XRD patterns of PdeCu membranes fabricated by
SIEP are shown in Fig. 7 at pre-HT and post-HT conditions. PreHT XRD spectra (Fig. 7a) show the reflection peaks of Pd and
Cu in the face centered cubic (f.c.c.) phase at {111}, {200}, {220}
and {311} planes. Table 5 summarizes the 2q and d-spacing
values corresponding to the four major reflection peaks for
pure Pd and PdeCu film for comparison [38,40]. The 2q values
suggest that all of the characteristics peaks of Pd are in the
same position for both Pd and PdeCu films. In addition, the
presence of different bravais lattices signify the formation of
polycrystalline structure throughout the film. It is worth
mentioning that in our XRD analysis, no reflection peaks for
Fe, Cr, Ni and Mn/Mo metals were observed. The PdeCu
membranes were annealed in a hydrogen environment at
773 K for 10 and 18 h to better understand the alloying process.
In Fig. 7b, after 10 h of HT we observe that Pd and Cu in {111},
{200}, and {220} planes are merged into three distinct peaks at
Pd planes. This shows formation of a PdeCu alloy. After 18 h of
HT, as shown in Fig. 7c, the three peaks become sharper. The
shifting of peaks indicates the coexistence of Pd and Cu in the
deposited film and formation of PdeCu interstitial alloys
[28,42,43]. Interstitial alloy formation occurs when the atoms
of one component are smaller than the other, and the smaller
atoms fit into the spaces (interstices) between the larger
atoms. During 18 h of annealing, we did not observe alloy
formation between Pd and Cu at {311} plane. Allowing more
time could result alloy formation at {311} planes.
The pores in the MPSS substrate are tortuous and interconnected. To understand the PdeCu deposition in the pores
and on the substrate surface, we examined the cross-section
of PdeCu SIEP membranes using EDS. As an illustrative
example, in Fig. 8a, we probed four locations across a depth of
25 mm for Pd and Cu metal contents, which suggests that we
were able to deposit metals deep inside the pores. The Pd and
Cu metal contents in these pores are shown in Fig. 8b by a bar
Fig. 8 e SEM images of PdeCu film cross-section: showing the locations of pores to study Pd and Cu metal distribution by
EDS analysis starting from the pore mouth to deep inside (from Probe 1 / Probe 4).
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 3 4 7 7 e3 4 9 0
3485
Fig. 9 e EDS line scanning of PdeCu film cross-section (scanning length 25 mm, scanning direction from (a) / (b)).
graph. The Cu content inside the pores decreases across the
depth, which is expected because of the sequential EP deposition of Cu over Pd. This conclusion is reaffirmed by the EDS
line scanning of the PdeCu film cross-section, as shown in
Fig. 9. It may be mentioned here that in this line scanning, we
did not observe Fe, Cr, and Ni metals in the PdeCu film. The
oscillation of X-ray counts was prominent for Cu, as the Cu
content is low in each deposition cycle. Being the minor
constituent, Ni has the background effect and oscillates more.
However, the sharp rise of peaks for Fe, Cr and Ni (sharp
trough for Pd and Cu) in the middle of the PdeCu alloy film
was observed. Indeed, these were the response from fine
stainless steel particles doped in the film during metal polishing while preparing the sample.
3.3.
Hydrogen permeability studies
The PdeCu membranes fabricated by the CEP and SIEP techniques were tested for hydrogen perm-selectivity at pre-HT
and post-HT conditions. As shown in Table 4 in the previous
section, the SIEP membrane samples are numbered as S1
through S4 while the CEP membranes are numbered as C1 and
C2. In this table, we summarized our hydrogen flux and
selectivity data, along with membrane thickness.
For dense metallic membranes (10 mm), it is believed that
the diffusion of atomic hydrogen through the metal films is
the rate-controlling step [44] and is described by SievertsFick’s law as:
JH2 ¼
QH n
PH pnH
t
(1)
where, QH is the hydrogen permeability (a product of solubility
and diffusivity), t is the membrane thickness or film thickness,
and PH and pH are the partial pressures of hydrogen on the
high and low pressure sides, respectively. The exponent n is
a parameter, whose value depends on the limiting transport
mechanism. If H2 surface reactions at the permeance side
limit the flux, then the power index n approaches to zero.
When diffusion through the membrane limits the hydrogen
flux, n becomes 1/2 due to the dissociation of molecular
hydrogen into atomic hydrogen. On the other hand, if the gas
resistance or surface reactions at the retentate side are rate
limiting in hydrogen transport, n approaches to unity
[7,45e50].
We tested several PdeCu membrane samples fabricated
by the SIEP and CEP methods at pre-HT and post-HT conditions for hydrogen permeability and analyzed the flux data
using Eqn. (1). We observed that the power index ranged
between 0.5 and 1 (Table 4). This suggests that the H2transport through PdeCu membranes not only dominated by
3486
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 3 4 7 7 e3 4 9 0
solution-diffusion mode of transport but the gas phase
resistance and/or surface reactions may be rate limiting. As
an illustrative example, in Fig. 10, we present hydrogen flux
and selectivity data of PdeCu membranes fabricated by SIEP
(S2) and CEP (C2) at pre-HT condition. The permeability
measurements were carried out in the temperature range of
523e823 K with transmembrane pressure of 20e100 psi. The
H2 flux data is presented as a function of ðPnH pnH Þ and from
a non-linear least-square analysis we find the power index n
as 0.85 for S2 (Fig. 10a) and 0.93 for C2 (Fig. 10b). The lower
value of the power index (n) of the membranes fabricated by
the SIEP method suggests that these membranes offer
hydrogen atoms more diffusion through bulk metals
compared to the membranes fabricated by CEP method. The
selectivity of hydrogen over nitrogen of the two membranes
is presented in Fig. 10c and d for S2 and C2 membrane,
respectively as a function of ðPnH pnH Þ. The selectivity is
defined as:
aH2 =N2 ¼
JH2
JN2
(2)
where, Ji is the measured flux of the subscripted gases. For
both of the membranes, hydrogen selectivity increases with
increasing temperature, whereas the selectivity decreases
with increasing pressure. However, the SIEP membrane (S2)
had higher permeability and selectivity compared to the CEP
membrane (C2). For the four PdeCu membranes (S2, S3, C1
and C2) that we tested for hydrogen perm-selectivity, we
observed highest selectivity at 823 K and 20 psi pressure in
pre-HT membranes. The measured selectivity was 84.1 for the
S2 membrane and 19.7 for the C2 membrane. The result shows
that the SIEP membrane provides significantly higher selectivity compared to the CEP membrane. In this example, the
SIEP membrane had about four-fold higher selectivity at 823 K
and 20 psi pressure than the CEP membrane.
The SIEP (S2 and S3) and CEP (C1 and C2) membranes were
heat-treated for 18 h in a hydrogen environment at 773 K.
a
b
c
d
Fig. 10 e Hydrogen flux and H2/N2 selectivity data of PdeCu membranes fabricated by SIEP and CEP methods (pre-HT).
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 3 4 7 7 e3 4 9 0
These membranes were then tested for hydrogen permselectively as before. The H2 flux and selectivity data of the
post-HT S2 and C2 membranes are presented in Fig. 11 as
a function of ðPnH pnH Þ. From flux data analysis, the power
index was found to be 0.80 for the S2 membrane (Fig. 11a). In
comparison, the power index was found to be 0.91 for the C2
membrane (Fig. 11). In both membranes, hydrogen flux
increased with temperature for a given transmembrane
pressure. However, if we compare with pre-HT membrane
fluxes, we find that the post-HT S2 membrane has significantly higher flux (Fig. 11a) compared to the pre-HT S2
membrane flux (Fig. 10a). In the case of CEP membranes, the
pre- and post-HT C2 membrane the flux gain was found to be
marginal (Figs. 10b and 11b). The H2-selectivity data of S2 and
C2 membranes, presented in Fig. 11c and d, show that the SIEP
membrane has significantly higher selectivity relative to CEP
membrane. If we compare the selectivity of the pre- and postHT S2 membrane, we find that the selectivity increased from
3487
84 to 186 at 20 psi and 823 K. For the CEP C2 membrane,
however, we did not observe any change in selectivity at the
pre- and post-HT conditions.
To illustrate the intrinsic membrane behavior of the SIEP
and CEP PdeCu membranes, we computed the permeability
coefficients QH at four different temperatures using an
Arrhenius plot (QH vs. 1/T ), shown in Fig. 12. The permeability
data fit very well (correlation coefficient, r2 ¼ 0.98 in both
cases) to the Arrhenius equation:
QH ¼ QHo expðE=RTÞ
(3)
where, QHo is the reference permeance, E is the activation
energy, T is the absolute temperature and R is the universal
gas constant. The E value of the SIEP PdeCu membrane (film
thickness of 16.73 mm) was found to be 9.4 kJ/mol, whereas the
corresponding E value of activation energy of the CEP PdeCu
membrane (film thickness of 20.17 mm) was 5.82 kJ/mol. In our
previous study we found the E values of the SIEP Pd membrane
a
b
c
d
Fig. 11 e Hydrogen flux and H2/N2 selectivity data of PdeCu membranes fabricated by SIEP and CEP methods (post-HT).
3488
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 3 4 7 7 e3 4 9 0
0.18
-0.6
Thermal Cycling @ 10 psi
Pd - Cu SIEP Membrane
Pd - Cu CEP Membrane
-0.8
0.16
873 K
-1.0
0.14
-1.2
-1.4
H2 Flux (mol/m2-s)
0.12
ln QH
-1.6
-1.8
-2.0
-2.2
0.10
0.08
573 K
0.06
-2.4
0.04
-2.6
0.02
-2.8
-3.0
0.0010
0.0012
0.0014
0.0016
0.0018
0.00
0.0020
0
10
20
-1
1/T (K )
30
40
50
60
70
80
90
100
Number of Days
Fig. 12 e Arrhenius plots of H2-permeability coefficients of
PdeCu membranes fabricated by SIEP and CEP methods.
Fig. 13 e H2 flux data of PdeCu MPSS membrane under
thermal cycling fabricated by SIEP method.
(film thickness of 7.68 mm) and SIEP PdeAg membranes (film
thickness of 12.63 mm) as 16.88 and 11.93 kJ/mol, respectively
[38,39]. These results are summarized in Table 6. We observed
that, with increasing membrane film thickness, the activation
energy decreases and consequently, the power index of Sieverts’ law increases. A low value of the activation energy
((30 kJ/mol) indicates that the surface phenomena of dissociative adsorption and recombinative desorption do not
significantly influence the permeation process since they are
characterized by a markedly higher activation energy
(w54e146 kJ/mol) [48]. The reported activation energy may be
viewed as “apparent activation energy” since the H2-permeation in these membranes is contributed to more than one
phenomenon.
The PdeCu membranes prepared by SIEP method were
subjected to thermal cycling to check their performance. For
a period of 90 days, H2 flux of the membrane was recorded
under thermal cycling of 573 Ke873 K e 573 K, at 10 psi pressure in on and off mode (8e10 h a day) in our permeability
measurement set-up. On each day of testing, the membrane
underwent a couple of thermal cycles, and the average data of
H2 flux for each day are presented in Fig. 13. The fluctuation in
the hydrogen flux is attributed to the PdeCu alloying in
progress (as discussed earlier in Fig. 7). In this thermal cycling
test, the membrane did not deteriorate and remained stable in
terms of hydrogen permeability. Our recorded average
H2fluxes were 0.0922 mol/m2-s (standard deviation of 1.27%)
and 0.13357 mol/m2-s (standard deviation of 1.65%) at 573 K
and 873 K, respectively.
Table 6 e Comparison of values of activation energy of
different membranes fabricated by SIEP and CEP
methods.
Membrane Fabrication Membrane Sieverts’ Activation
film
method
film
law
energy, E
thickness
power
(kJ/mol)
(mm)
index, n
Pd [39]
PdeAg [38]
PdeCu
PdeCu
SIEP
SIEP
SIEP
CEP
7.68
12.63
16.73
20.17
0.61
0.75
0.8
0.91
16.88
11.93
9.4
5.82
4.
Conclusions
Using selected surfactant in SIEP, a defect-free PdeCu
membrane on MPSS support was fabricated. The SIEP
membrane thickness was relatively thinner compared to the
CEP PdeCu membrane with about 14.4% reduced fabrication
time. From SEM analysis, it was found that SIEP PdeCu
membranes had finer grain structure relative to CEP
membranes. The PdeCu membrane fabricated by SIEP method
had a smoother surface structure compared to the CEP
membrane, as confirmed by AFM analysis. Upon heat treatment, excellent grain agglomeration was observed in the SIEP
membrane with significant grain fusion. Cross-sectional EDS
analysis of the SIEP PdeCu membrane shows deep penetration of metals in the pores of up to 25 mm. XRD spectra
confirmed the polycrystalline structure of the PdeCu film after
18 h of annealing at 773 K under a hydrogen environment.
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 3 4 7 7 e3 4 9 0
The membranes fabricated by the SIEP and CEP methods
displayed significantly different hydrogen transport behaviors
after heat treatment. The H2 flux and selectivity of the CEP
PdeCu membrane were found to be insensitive to annealing
process. On the other hand, the SIEP PdeCu membrane
showed remarkable flux and selectivity enhancement upon
heat treatment. The long-term (for a period of 90 days)
thermal stability data demonstrated that the SIEP PdeCu
membrane was very stable in rigorous exposure to low and
high temperature thermal cycling.
Acknowledgments
This research was sponsored by the U.S. Department of
Energy e HBCU Program, under Award No. DE-FG08NT000143. Mr. Richard Dunst, NETL, Morgantown, is the DOE
Project Officer. However, any opinions, findings, and conclusions or recommendations expressed herein are those of the
authors and do not necessarily reflect the views of the DOE.
Analytical support from the Center for Advanced Materials
and Smart Structures (CAMSS) of North Carolina A&T State
University is gratefully acknowledged.
[11]
[12]
[13]
[14]
[15]
[16]
[17]
references
[18]
[1] Xuan J, Leung MKH, Leung DYC, Ni M. A review of biomassderived fuel processors for fuel cell systems. Renewable and
Sustainable Energy Reviews 2009;13:1301e13.
[2] Thoen PM, Roa F, Way JD. High flux palladiumecopper
composite membranes for hydrogen separations.
Desalination 2006;193:224e9.
[3] Roa F, Block MJ, Way JD. The influence of alloy composition
on the H2 flux of composite PdeCu membranes. Desalination
2002;147:411e6.
[4] Pomerantz N, Ma YH. Novel method for producing high H2
permeability Pd membranes with a thin layer of the sulfur
tolerant Pd/Cu fcc phase. Journal of Membrane Science 2011;
370:97e108.
[5] Gao H, Lin JYS, Li Y, Zhang B. Electroless plating
synthesis, characterization and permeation properties of
PdeCu membranes supported on ZrO2 modified porous
stainless steel. Journal of Membrane Science 2005;265:
142e52.
[6] Yang JY, Nishimura C, Komaki M. Effect of overlayer
composition on hydrogen permeation of PdeCu alloy coated
V-15Ni composite membrane. Journal of Membrane Science
2006;282:337e41.
[7] Yang JY, Komaki M, Nishimura C. Effect of overlayer
thickness on hydrogen permeation of Pd60Cu40/V-15Ni
composite membranes. International Journal of Hydrogen
Energy 2007;32:1820e4.
[8] Hoang HT, Tong HD, Gielens FC, Jansen HV, Elwenspoek MC.
Fabrication and characterization of dual sputtered PdeCu
alloy films for hydrogen separation membranes. Materials
Letters 2004;58:525e8.
[9] Yang JY, Nishimura C, Komaki M. Hydrogen permeation of
Pd60Cu40 alloy covered V-15Ni composite membrane in
mixed gases containing H2S. Journal of Membrane Science
2008;309:246e50.
[10] Ma YH, Akis BC, Ayturk ME, Guazzone F, Engwall EE,
Mardilovich IP. Characterization of intermetallic diffusion
barrier and alloy formation for Pd/Cu and Pd/Ag porous
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
3489
stainless steel composite membranes. Industrial &
Engineering Chemistry Research 2004;43:2936e45.
Chen W-H, Hsu P-C. Hydrogen permeation measurements of
Pd and PdeCu membranes using dynamic pressure
difference method. International Journal of Hydrogen Energy
2011;36:9355e66.
Ward TL, Dao T. Model of hydrogen permeation behavior in
palladium membranes. Journal of Membrane Science 1999;
153:211e31.
Ayturk ME, Mardilovich IP, Engwall EE, Ma YH. Synthesis of
composite Pd-porous stainless steel (PSS) membranes with
a Pd/Ag intermetallic diffusion barrier. Journal of Membrane
Science 2006;285:385e94.
Yeung KL, Christiansen SC, Varma A. Palladium composite
membranes by electroless plating technique: relationships
between plating kinetics, film microstructure and
membrane performance. Journal of Membrane Science
1999;159:107e22.
Nam S-E, Lee K-H. Preparation and characterization of
palladium alloy composite membranes with a diffusion
barrier for hydrogen separation. Industrial & Engineering
Chemistry Research 2005;44:100e5.
Kim D-W, Park YJ, Moon J-W, Ryi S-K, Park J-S. The effect of
Cu reflow on the PdeCu-Ni ternary alloy membrane
fabrication for infinite hydrogen separation. Thin Solid Films
2008;516:3036e44.
Gao H, Lin YS, Li Y, Zhang B. Chemical stability and its
improvement of palladium-based metallic membranes.
Industrial & Engineering Chemistry Research 2004;43:
6920e30.
Li X, Liu TM, Huang D, Fan YQ, Xu NP. Preparation and
characterization of ultrathin palladium membranes.
Industrial & Engineering Chemistry Research 2009;48:2061e5.
Knapton AG. Palladium alloys for hydrogen diffusion
membranes. Platinum Metals Review 1977;21:44e50.
Nam S-E, Lee K-H. Hydrogen separation by Pd alloy
composite membranes: introduction of diffusion barrier.
Journal of Membrane Science 2001;192:177e85.
Grashoff GJ, Pilkington CE, Corti CW. The purification of
hydrogen. Platinum Metals Review 1983;27:157e69.
Mckinley DL. Metal alloy for hydrogen separation and
purification. United States Patent 1967;3350845.
Bosko ML, Yepes D, Irusta S, Eloy P, Ruiz P, Lombardo EA,
Cornaglia LM. Characterization of PdeAg membranes after
exposure to hydrogen flux at high temperatures. Journal of
Membrane Science 2007;306:56e65.
Ayturk ME, Ma YH. Electroless Pd and Ag deposition kinetics
of the composite Pd and Pd/Ag membranes synthesized from
agitated plating baths. Journal of Membrane Science 2009;
330:233e45.
Basile A, Gallucci F, Iulianelli A, De Falco M, Liguori S.
Hydrogen production by ethanol steam reforming:
experimental study of a PdeAg membrane reactor and
traditional reactor behaviour. International Journal of
Chemical Reactor Engineering 2008;6:1e16.
Huang T-C, Wei M-C, Chen H-I. Preparation of
palladiumesilver alloy composite membranes for hydrogen
permeation. Chemical Engineering Communications 2002;
189:1262e82.
Nair BKR, Choi J, Harold MP. Electroless plating and
permeation features of Pd and Pd/Ag hollow fiber composite
membranes. Journal of Membrane Science 2007;288:67e84.
Roa F, Way JD, McCormick RL, Paglieri SN. Preparation and
characterization of PdeCu composite membranes for
hydrogen separation. Chemical Engineering Journal 2003;93:
11e22.
Gallucci F, De Falco M, Tosti S, Marrelli L, Basile A. Co-current
and counter-current configurations for ethanol steam
3490
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 3 4 7 7 e3 4 9 0
reforming in a dense PdeAg membrane reactor.
International Journal of Hydrogen Energy 2008;33:6165e71.
Tierney HL, Baber AE, Sykes ECH. Atomic-scale imaging and
electronic structure determination of catalytic sites on Pd/Cu
near surface alloys. The Journal of Physical Chemistry C
2009;113:7246e50.
Yang JY, Nishimura C, Komaki M. Preparation and
characterization of PdeCu/V-15Ni composite membrane for
hydrogen permeation. Journal of Alloys and Compounds
2007;431:180e4.
Islam MA, Ilias S. Characterization of Pd-composite
membrane fabricated by surfactant induced electroless
plating (SIEP): effect of grain size on hydrogen
permeability. Separation Science and Technology 2010;45:
1886e93.
Kulprathipanja A, Alptekin GO, Falconer JL, Way JD. Effects of
water gas shift gases on PdCu alloy membrane surface
morphology and separation properties. Industrial &
Engineering Chemistry Research 2004;43:4188e98.
Chen BH, Hong L, Ma Y, Ko TM. Effects of surfactants in an
electroless nickel-plating bath on the properties of NiP
alloy deposits. Industrial & Engineering Chemistry Research
2002;41:2668e78.
Souleimanova RS, Mukasyan AS, Varma A. Effects of osmosis
on microstructure of Pd-composite membranes synthesized
by electroless plating technique. Journal of Membrane
Science 2000;166:249e57.
Ilias S, Islam MA. Methods of preparing thin films by
electroless plating. United States Patent; 2010.
Huang T-C, Wei M-C, Chen H-I. Preparation of hydrogenpermselective palladiumesilver alloy composite membranes
by electroless co-deposition. Separation and Purification
Technology 2003;32:239e45.
Rahman MM. "Fabrication of Pd and PdeAg membranes by
surfactant induced electroless plating (SIEP)," MS thesis,
chemical and Bioengineering, North Carolina A & T State
University, Greensboro, NC, 2010.
Islam MA. "The development of improved electroless plating
in fabricating Pdebased membrane and membrane reactor
application for hydrogen separation," PhD Dissertation,
[40]
[41]
[42]
[43]
[44]
[45]
[46]
[47]
[48]
[49]
[50]
Energy and Environmental Studies, North Carolina A & T
State University, Greensboro, NC, 2008.
Islam MS "Fabrication of PdeCu membranes by surfactant
induced electroless plating (SIEP)," MS thesis, Chemical and
Bioengineering, North Carolina A&T State University,
Greensboro, 2011.
Manne S, Gaub HE. Molecular organization of surfactants at
solid-liquid interfaces. Science 1995;270:1480e2.
Mardilovich PP, She Y, Ma YH, Rei M-H. Defect-free
palladium membranes on porous stainless-steel support.
AIChE Journal 1998;44:310e22.
Lin W-H, Chang H-F. Characterizations of PdeAg membrane
prepared by sequential electroless deposition. Surface and
Coatings Technology 2005;194:157e66.
Roa F, Way JD. Influence of alloy composition and membrane
fabrication on the pressure dependence of the hydrogen flux
of palladiumcopper membranes. Industrial & Engineering
Chemistry Research 2003;42:5827e35.
Gielens FC, Tong HD, Vorstman MAG, Keurentjes JTF.
Measurement and modeling of hydrogen transport through
high-flux Pd membranes. Journal of Membrane Science 2007;
289:15e25.
Collins JP, Way JD. Preparation and characterization of
a composite palladium-ceramic membrane. Industrial &
Engineering Chemistry Research 1993;32:3006e13.
Uemiya S, Sato N, Ando H, Kude Y, Matsuda T, Kikuchi E.
Separation of hydrogen through palladium thin film
supported on a porous glass tube. Journal of Membrane
Science 1991;56:303e13.
Caravella A, Scura F, Barbieri G, Drioli E. Sieverts law
Empirical exponent for Pd-based membranes: critical
analysis in pure H2 permeation. The Journal of Physical
Chemistry B 2010;114:6033e47.
Caravella A, Barbieri G, Drioli E. Modelling and simulation of
hydrogen permeation through supported Pd-alloy
membranes with a multicomponent approach. Chemical
Engineering Science 2008;63:2149e60.
Chen W-H, Chiu IH. Transient dynamic of hydrogen
permeation through a palladium membrane. International
Journal of Hydrogen Energy 2009;34:2440e8.