Influence of Nucleating Agents on the Chemical Interaction of MgO-Al 2O3-SiO2-B2O3 Glass Sealants with Components of SOFCs

Article abstract of DOI:10.1149/1.1647570

The chemical interactions between different MgO-SiO₂-Al ₂O₃-B₂O₃ base glass sealants and components of solid oxide fuel cells (SOFCs) were investigated. The SOFC materials considered are ZrO₂ stabilized with 8 mol % Y ₂O₃ (as electrolyte and part of the anode), Ni (as part of the anode), and the oxide-dispersion-strengthened (ODS) alloy Cr5Fel Y ₂O₃ (as interconnect). Glass compositions with no nucleating agent and with ZrO₂, Cr₂O₃, or Ni as the nucleating agents were prepared. Powder mixtures of these sealants with the mentioned SOFC materials, as well as the sealant/material interfaces, were characterized by X-ray diffraction and scanning electron microscopy and energy dispersive spectroscopy in order to determine possible reaction phases and the diffusion behavior of different cations. Formation of cordierite and cristobalite as detrimental phases was detected in many of the mixtures. The formation of these phases can be suppressed if Cr₂O₃ or Ni is added to the glass as the nucleating agent. The most interesting feature of these results is the absence of the cordierite phase for all reaction powder mixtures if Cr₂O₃ is used as the nucleating agent. The sealants with Cr₂O₃ and Ni as nucleating agents formed a reaction zone at the interface with ODS, rich in Cr and Mg. A parabolic reaction rate equation describing the growth of the reaction zone thickness, the diffusion coefficient of chromium, and rate constants was evaluated.

Full text of DOI:10.1149/1.1647570

A558
Journal of The Electrochemical Society, 151 ͑4͒ A558-A562 ͑2004͒
0013-4651/2004/151͑4͒/A558/5/$7.00 © The Electrochemical Society, Inc.
Influence of Nucleating Agents on the Chemical Interaction
of MgO-Al2O3-SiO2-B2O3 Glass Sealants with Components
of SOFCs
D. Bahadur,a N. Lahl, K. Singh, L. Singheiser, and K. Hilpertz
¨
¨
Institute for Materials and Processing in Energy Systems, Forschungszentrum Julich GmbH, 52425 Julich,
Germany
The chemical interactions between different MgO-SiO2-Al2O3-B2O3 base glass sealants and components of solid oxide fuel cells
͑SOFCs͒ were investigated. The SOFC materials considered are ZrO2 stabilized with 8 mol % Y2O3 ͑as electrolyte and part of the
anode͒, Ni ͑as part of the anode͒, and the oxide-dispersion-strengthened ͑ODS͒ alloy Cr5Fe1Y2O3 ͑as interconnect͒. Glass
compositions with no nucleating agent and with ZrO2 , Cr2O3 , or Ni as the nucleating agents were prepared. Powder mixtures of
these sealants with the mentioned SOFC materials, as well as the sealant/material interfaces, were characterized by X-ray
diffraction and scanning electron microscopy and energy dispersive spectroscopy in order to determine possible reaction phases
and the diffusion behavior of different cations. Formation of cordierite and cristobalite as detrimental phases was detected in many
of the mixtures. The formation of these phases can be suppressed if Cr2O3 or Ni is added to the glass as the nucleating agent. The
most interesting feature of these results is the absence of the cordierite phase for all reaction powder mixtures if Cr2O3 is used as
the nucleating agent. The sealants with Cr2O3 and Ni as nucleating agents formed a reaction zone at the interface with ODS, rich
in Cr and Mg. A parabolic reaction rate equation describing the growth of the reaction zone thickness, the diffusion coefficient of
chromium, and rate constants was evaluated.
© 2004 The Electrochemical Society. ͓DOI: 10.1149/1.1647570͔ All rights reserved.
Manuscript submitted March 31, 2003; revised manuscript received October 5, 2003. Available electronically February 20, 2004.
One of the technological problems in the development of planar
3. The powder mixtures of most magnesium aluminosilicate
sealants with the ODS alloy or the steel Fe18Cr1Al ͑DIN 1.4742͒
did not show the lines of the cordierite phase in the X-ray diffraction
͑XRD͒ pattern. However, the presence of the cristobalite phase is
seen in many cases and there seems to be a competition between the
formation of cordierite and cristobalite. It may be mentioned that the
diffusion of chromium into the sealant is the highest for the ODS
alloy and the steel.
solid oxide fuel cells ͑SOFCs͒ is the sealing of the stack.1-4 Among
the many glasses investigated for this purpose, aluminosilicate glass
is a good candidate.
We
recently
investigated
a
large
number
of
AO-SiO2-Al2O3-B2O3 (A ϭ Ba, Ca, Mg) base glasses. By de-
tailed microstructural and structural studies and by thermal analysis,
we tried to understand their crystallization and thermal expansion
behavior.5,6 Powder mixtures and diffusion couples were investi-
gated by us7 in order to elucidate the chemical interactions at the
interface between sealants and other materials of SOFC at the anode
side. These materials are ZrO2 stabilized with 8 mol % Y2O3
͑8YSZ͒ as the electrolyte and part of the anode, Ni as part of the
anode, and the oxide-dispersion-strengthened ͑ODS͒ alloy
Cr5Fe1Y2O3 Ducrolloy ͑supplied by Plansee, Reute, Austria͒ as
interconnect. It was observed that MgO base sealants are superior to
the BaO or CaO base sealants because of a higher activation energy
of crystal growth thereby hindering crystallization and moderate
chemical interactions at the interface. However, on interacting with
different components of SOFC MgO base aluminosilicate glasses
often lead to the formation of the two detrimental phases cordierite
and cristobalite. The thermal expansion coefficient ͑TEC͒ of the
cordierite phase is very low (2 ϫ 10Ϫ6/K) compared to that of the
other components of the fuel cell ͑FC͒ which is detrimental to the
SOFC. Cristobalite is detrimental because it undergoes a structural
transformation at 200°C involving a large volume change, which
causes microcracking during cooling from the sealing temperature.
The following features of the results on several magnesium alumi-
nosilicate glasses7 indicate the need for further work on the devel-
opment of such glasses.
In view of the above noteworthy features, we chose three MgO
base aluminosilicate glasses with ZrO2 , Ni, or Cr2O3 as the nucle-
ating agents, respectively. The high covalent nature of the bonding
of Ni and Cr ions and their high field strength values can enhance
the rearrangement of the glass structure and modify the glass tran-
sition and crystallization behavior. Indeed, the glass transition tem-
perature, Tg , and the crystallization temperature, Tc , of the two
glasses with nickel and chromia as nucleating agents were the high-
est of all the sealants investigated so far by us.5,6 They also exhib-
ited high values for the activation energy of crystal growth.5 For
example, the values of the activation energy for the sealant without
and with ZrO2 as nucleating agent amount to 420 and 410 kJ/mol,
respectively, whereas these values are 498 and 622 kJ/mol if chro-
mia and nickel are used as nucleating agents, respectively. A high
stability and compatibility have, therefore, to be expected for these
two sealants. Nucleating agents play a significant role in the homo-
geneous nucleation and growth of phases. The growth of selected
phases may be favored while that of other phases is restricted or
suppressed. The choice of nucleating agents and the glass system in
the present studies is based on the results obtained in our previous
reports.5-7 We present here the results of their chemical interactions
with 8YSZ ͑as electrolyte and part of the anode͒, Ni ͑as part of the
anode͒ and the ODS alloy ͑as interconnect͒. Moreover, the diffusion
behavior of different cations, in particular chromium, has been in-
vestigated. We also tried to explore the reasons for the suppression
of undesirable phases.
1. When a particular MgO-Al2O3-SiO2 base sealant and its
powder mixture with nickel were annealed under similar conditions,
the cordierite phase was completely suppressed in the latter.
2. When a MgO-Al2O3-SiO2 base sealant with TiO2 as nucleat-
ing agent and its powder mixture with 8YSZ were annealed under
similar conditions, again the cordierite phase was only suppressed in
the latter.
Experimental
The glass compositions considered here together with their label
are given in Table I. The details of the synthesis of glasses and their
characterization are given elsewhere.5,6 The differential thermal
analysis ͑DTA͒ studies were carried out using the thermal analyzer-
type STA 429 supplied by Netzsch, Selb, Germany. The chemical
interactions between the sealants and the components of the SOFC
were investigated by annealing their powder mixtures and diffusion
a Permanent address: Department of Metallurgical Engineering and Materials
Science, Indian Institute of Technology Bombay, Mumbai 400076, India.
z E-mail: k.hilpert@fz-juelich.de
Downloaded on 2015-07-10 to IP 128.210.126.199 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 151 ͑4͒ A558-A562 ͑2004͒
A559
Table I. Composition of glasses investigated in mole percent with
label.
Glass MgO
Ni
Label
SiO2 Al2O3 B2O3 ZrO2 Cr2O3
1
2
3
4
45
38
45
45
5
10
6.4
6.6
5
5
3.4
3.7
MAS
MASZ10
1,9 MASN6
MASC6
2
43.7 44.6
46.2 42.9
0,6
couples under ambient air. Phase analysis was conducted by XRD
using Cu K␣ and Co K␣ radiation ͑Philips 1050, Eindhoven, The
Netherlands͒. The chemical composition of the phases was obtained
by scanning electron microscopy ͑SEM͒ with energy-dispersive
X-ray analysis ͑EDX, CAMSCAN/Tracor Northern, Middleton,
WI͒. SEM/EDX was used to determine the diffusion profile of dif-
ferent cations at the interface of the diffusion couples.
Figure 2. Plot of activation energy of crystal growth vs. electronegativity of
cations used as nucleating agents.
Results and Discussion
have tetrahedral coordination in addition to the higher coordination.
In contrast to this, the Zr4ϩ ion can only have higher coordination
and cannot act as a network former.
Figure 1 shows the DTA curves determined at a heating rate of
10 K per min for the samples MASN6 and MASC6. The DTA re-
sults of the other glasses investigated by us are reported in Ref. 5.
The glass transition temperature, Tg , and the crystallization tem-
perature, Tc , are determined by DTA. The values for Tg are 717,
723, 735, and 738°C for the glasses MAS, MASZ10, MASN6, and
MASC6, respectively, whereas the Tc values are 921, 931, 939, and
946, respectively. The Tg and Tc values are significantly higher for
the samples with nucleating agents. Activation energies of crystal
growth were determined from the modified Kissinger equation as
described in our earlier paper.5 The values of the activation energy
are very high for the glasses containing Ni or Cr2O3 as the nucleat-
ing agent. The value of the activation energy for the glass MAS with
no nucleating agent is 420 kJ/mol. Figure 2 shows the plot of the
activation energy vs. the electronegativity for the different ions used
as nucleating agents. The electronegativity which determines the
bonding character seems to have a marked influence on the activa-
tion energy, though the concentration of the nucleating agents in the
glass is very small. However, this is merely an observed correlation.
All nucleating ions considered in this work, except Zr4ϩ, enhance
the activation energy of crystal growth. This is possibly due to the
fact that these ions may act as network formers because they can
Chemical interaction in powder mixtures.—Chemical interac-
tions were investigated by annealing powder mixtures made of the
different sealants and Ni, the ODS alloy, or 8YZS at 1000°C for 100
h. Results on the subsequent characterization of the mixtures are
given in Table II. The results showed no interesting features if ZrO2
was used as the nucleating agent. We therefore limit our further
discussion to the results obtained with two sealants containing
Cr2O3 or Ni as nucleating agents.
Figure 3 shows the XRD pattern of the powder mixtures MAS
ϩ Ni, MASN6 ϩ Ni, and MASC6 ϩ Ni annealed in air at
1000°C for 500 h. The pattern indicates the presence of the phases
Mg2SiO4 ͑forsterite͒ and MgSiO3 . The cordierite phase
(Mg2Al4Si5O18) is evident in the powder mixtures of MAS ϩ Ni
and MASN6 ϩ Ni. The exception is the mixture MASC6 ϩ Ni
where Cr2O3 is used as the nucleating agent. It is very encouraging
to note that the undesirable cordierite phase is completely sup-
pressed in this powder mixture. There is another interesting obser-
vation with regard to the product phases: The lines of the MgSiO3
product phase show high intensity for the powder mixture of MAS
with Ni, whereas the lines of the Mg2SiO4 product phase show high
intensity for the other two mixtures where nucleating agents are
used. As mentioned above, our DTA results show a significantly
higher Tg and Tc for the sealants with the two nucleating agents.
This appears to change the crystallization kinetics and the formation
of product phases in the powder mixtures. We observed a similar
change in the volume fraction of the MgSiO3 and Mg2SiO4 phases
with varying amounts of TiO2 as the nucleating agent in
MgO-Al2O3-SiO2-B2O3 base glasses.7
Figure 4 shows the XRD pattern of the powder mixtures MAS
ϩ 8YSZ, MASN6 ϩ 8YSZ, and MASC6 ϩ 8YSZ annealed in
air at 1000°C for 100 h. The results are very similar to Fig. 3 with
the exception that the additional phase ZrSiO4 was detected in all
samples as the dominant product phase. Moreover, the product
phases Mg2SiO4 and MgSiO3 are present in the powder mixtures.
The ratio of the volume fraction of MgSiO3 and Mg2SiO4 changes
in a similar way to that observed in the case of the powder mixtures
with nickel. The formation of the cordierite phase is suppressed in
the powder mixture MASC6 ϩ 8YSZ. Note that the suppression of
the cordierite phase was also reported by us for a mixture of a
MgO-Al2O3-SiO2-B2O3 base sealant with 8YSZ, if TiO2 was used
as the nucleating agent.7
Figure 5 shows the XRD pattern of the powder mixtures MAS
ϩ ODS, MASN6 ϩ ODS, and MASC6 ϩ ODS annealed in air at
1000°C for 500 h. The XRD patterns are nearly identical. The com-
Figure 1. DTA thermograms determined at a heating rate of 10 K minϪ1 for
samples MASN6 and MASC6.
Downloaded on 2015-07-10 to IP 128.210.126.199 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A560
Journal of The Electrochemical Society, 151 ͑4͒ A558-A562 ͑2004͒
Table II. Reaction products determined by XRD in powder mixtures composed of different sealants and SOFC materials after annealing at
1000°C for 100 h in air.
SOFC materials
Sealants
MAS
Ni
ODS
YSZ
Mg2Al4Si5O18 ,Mg2SiO4 ,
NiO,MgSiO3
(Mg,Fe)2SiO4 ,ZrSiO4 ,
MgSiO3 ,SiO2,
(Mg,Fe͒͑Cr,Al)2O4
Mg2SiO4 ,ZrSiO4 ,
Mg2Al4Si5O18
MgSiO3
,
MASZ10
Mg2Al4Si5O18 ,Mg2SiO4 ,
ZrSiO4 ,NiO,MgSiO3
(Mg,Fe)2SiO4 ,ZrSiO4 ,
Mg2Al4Si5O18 ,MgSiO3,
(Mg,Fe͒͑Cr,Al)2O4
Mg2SiO4 ,ZrSiO4 ,
Mg2Al4Si5O18
MgSiO3
,
MASN6
MASC6
Mg2Al4Si5O18 ,Mg2SiO4 ,
Mg0.4Ni0.6O,MgSiO3
Mg2SiO4 ,Mg0.4Ni0.6O,
MgAl2O4
Fe2SiO4 ,MgSiO3 ,SiO2,
(Mg,Fe͒͑Cr,Al)2O4
SiO2 ,Fe2SiO4 ,MgSiO3 ,
(Mg,Fe͒͑Cr,Al)2O4
ZrSiO4 ,Mg2SiO4 ,
Mg2Al4Si5O18
ZrSiO4 ,Mg2SiO4 ,
MgSiO3 ,MgAl2O4
phase separation is presumably responsible for the restriction in the
growth of cordierite phase in the powder mixtures. The Tg and Tc of
this glass is the highest of all the sealants investigated by us.5,6
Moreover, the activation energy of crystal growth is very high ͑498
kJ/mol͒. These observations suggest that the crystallization kinetics
of this glass is different from the other glasses. It may be noted that
the difference in the glass compositions seems to be too small to be
attributed to thermodynamic instability of the cordierite phase.
mon phases are MgSiO3 , spinel, and cristobalite. It is interesting to
note that the cordierite phase is absent in the powder mixtures. How-
ever, another detrimental phase, cristobalite, was detected. It may be
recalled that the cordierite phase was detected when the sealant
MAS was annealed at 1000°C for 1000 h, while the cristobalite
phase was absent.5 There seems to be some competition between the
formation of cordierite and cristobalite as discussed in our previous
paper.7
The most encouraging and interesting result is the complete ab-
sence of cordierite phase in all reaction mixtures of sealant MASC6
with the three SOFC materials. This is discussed in the following.
In our previous paper,7 we observed the suppression of the cordi-
erite phase in some reaction mixtures if TiO2 was used as nucleating
agent. This could be understood in terms of phase separation in-
duced by TiO2 . In the case of chromium oxide as the nucleating
agent, chromium can exist in the glass as 3ϩ and 6ϩ states if the
glass is melted under oxidizing conditions. The relative proportion
of these ions depends on the melting conditions or the other cations
present in the glass.8 Cr6ϩ represents an ion with a very high field
strength value of 2.40 compared to those of Ti4ϩ and Ni2ϩ, which
amount to 1.25 and 0.61, respectively. Moreover, it normally occu-
pies an interstitial position due to the small ionic radius of ϳ0.052
nm, which exerts a marked ordering effect upon the oxygen ions
surrounding it. Under these circumstances, a chromium-rich phase
can separate out from this glass.8 This is supported by our studies on
the diffusion couples that are reported in the following section. The
Investigation of diffusion couples.—A detailed investigation of
the diffusion couples MASN6/ODS and MASC6/ODS was carried
out after annealing at 1000°C for different time periods to study the
diffusion behavior of different cations across the interfaces. Figure 6
shows the SEM micrographs of these two diffusion couples an-
nealed for 22 h and 100 h, respectively, as typical examples. Figure
7 allows the comparison of the diffusion profiles of different cations
in the two diffusion couples. The reaction zone grows faster if Ni is
used as the nucleating agent in the sealant and shows major concen-
trations of Cr and Mg. The diffusion of Si in the reaction zone is
essentially negligible. The reasons for the formation of a Cr-Mg-rich
zone are discussed in the previous section as well as in our previous
paper.7 The reaction kinetics between Cr and Mg ion is rapid and
this favors the formation of the spinel phase. In such a situation, the
formation of the cordierite phase is inhibited. At the reaction tem-
Figure 3. XRD pattern of powder mixtures of ͑a͒ MAS ϩ Ni, ͑b͒
MASN6 ϩ Ni, and ͑c͒ MASC6 ϩ Ni after annealing in air at 1000°C for
500 h.
Figure 4. XRD pattern of powder mixtures of ͑a͒ MAS ϩ 8YSZ, ͑b͒
MASN6 ϩ 8YSZ, and ͑c͒ MASC6 ϩ 8YSZ after annealing in air at
1000°C for 500 h.
Downloaded on 2015-07-10 to IP 128.210.126.199 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 151 ͑4͒ A558-A562 ͑2004͒
A561
Figure 5. XRD pattern of powder mixtures of ͑a͒ MAS ϩ ODS, ͑b͒
MASN6 ϩ ODS, and ͑c͒ MASC6 ϩ ODS after annealing in air at 1000°C
for 500 h.
Figure 7. Typical diffusion profile of different cations at the interface of ͑a͒
MASN6 ϩ ODS, and ͑b͒ MASN6 ϩ ODS after annealing at 1000°C for
100 h or 22 h, respectively.
perature and in the reaction zone, mobile oxygen ions are always
available for the oxidation of chromium and formation of spinel
phase, thus maintaining the charge neutrality.
Figure 8 shows the square of the interdiffusion layer thickness,
X2, plotted vs. the diffusion time, t, for the interfaces of the sealants
MASN6 and MASC6 with the ODS alloy. The thickness of the
interdiffusion layer initially increases parabolically with time and
then becomes constant in practical terms after 20 h. The linear rela-
tion between X2 and t observed in the initial portion of the curves
strongly suggests that the formation of the interdiffusion layer is
Figure 8. Plot of square of thickness of reaction layer, X2, vs. time, t, for the
diffusion couples ͑᭹͒ MASN6/ODS and ͑᭿͒ MASC6/ODS after annealing
at 1000°C for different time periods.
Figure 6. Typical SEM micrographs of the interface of ͑a͒ MASC6
ϩ ODS, and ͑b͒ MASN6 ϩ ODS after annealing at 1000°C for 100 h or 22
h, respectively.
Downloaded on 2015-07-10 to IP 128.210.126.199 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A562
Journal of The Electrochemical Society, 151 ͑4͒ A558-A562 ͑2004͒
atom % of Cr in the sealant because the sealant contained Cr2O3 .
Table III. Diffusion coefficients, D, of chromium determined
The diffusion coefficients were determined by annealing the diffu-
sion couples at 1000°C for different time periods ͑Table III͒. The
average chemical diffusion coefficients for chromium at 1000°C for
the diffusion couples MASC6/ODS and MASN6/ODS resulted as
1.9 ϫ 10Ϫ12 cm2/s and 1.5 ϫ 10Ϫ12 cm2/s, respectively. This is in
good agreement with the value of the diffusion coefficient of Cr,
1.2 ϫ 10Ϫ12 cm2/s, which was obtained after annealing a diffusion
couple between the ODS alloy and a sealant containing TiO2 as
nucleating agent at 1000°C for 50 h.7
from the diffusion couples of MASN6ÕODS and MASZ6ÕODS af-
ter annealing at 1000°C for different time periods ⌬t. X is the
diffusion distance „see text….
Mean value of
Diffusion
couple
⌬t/h
X/␮m
D/(1012 cm2/s͒
D/(1012 cm2/s͒
MASN6/ODS
2
6
10
4
7
7
9
9
36
3
8
8
9
2.7
2.8
1.7
1.3
0.3
0.4
3.1
4.1
2.5
1.6
0.14
0.01
22
100
1100
2
6
10
20
100
1100
1.5
1.9
Conclusions
MASC6/ODS
The chemical interaction between MgO-SiO2-Al2O3-B2O3 base
glass sealants containing Cr2O3 , Ni, or ZrO2 as nucleating agents
with 8YSZ, Ni, and the ODS alloy were investigated. The most
encouraging result is the complete absence of the cordierite phase in
all powder mixtures composed of the sealant with Cr2O3 as the
nucleating agent and the FC materials. We attribute this to the high
values of Tg and Tc , and the activation energy of crystal growth if
Cr2O3 ͑0.6 mol %͒ is used as the nucleating agent. Moreover, chro-
mium seems to have a thermodynamic affinity to magnesium which
favors spinel formation and suppresses the formation of the cordier-
ite phase in many powder mixtures. Chemical diffusion coefficients
for the chromium cation were determined from the diffusion profile
and compared with literature values. The thickness of the interdif-
fusion layer initially increases parabolically with time and then be-
comes constant in practical terms after 20 h. The parabolic growth
suggests that the formation of the interdiffusion layer is controlled
by the diffusion of ions. The saturation behavior is understood in
terms of the changing composition of components with time.
6
10
controlled by the diffusion of ions. By using least squares computa-
tions, the linear portions in Fig. 8 can be described by the parabolic
reaction rate equation9
X2 ϭ kt
͓1͔
where k is the parabolic growth rate constant. The parabolic growth
rate constants evaluated are 3.7 ϫ 10Ϫ12 and 8.0 ϫ 10Ϫ12 cm2/s for
the sealants MASC6 and MASN6, respectively. The saturation be-
havior observed after 20 h can be understood in terms of the chang-
ing composition of the sealants. With time, the reaction layer be-
comes richer in chromium ͑from the ODS alloy͒ and in magnesium
͑from the sealant͒. Consequently, the concentration of magnesium in
the sealant and of chromium in the ODS alloy is depleted. The
depletion of magnesium ͑MgO is a network modifier͒ from the seal-
ant can raise the viscosity of the sealant and distort the open SiO4
tetrahedra. This can restrict the movement of ions, which may result
in the observed slowing down of the growth of the reaction zone.
Chemical diffusion coefficients were determined from the diffu-
sion profile of the various cations by using the relation10
The Institute for Materials and Processing in Energy Systems assisted in
meeting the publication costs of this article.
References
1. K. L. Ley, M. Krumpelt, R. Kumar, J. H. Meisser, and I. Bloom, J. Mater. Res., 11,
1489 ͑1996͒.
2. Y. M. Sung, J. Mater. Sci., 31, 5421 ͑1996͒.
¨
3. C. Gunther, G. Hofer, and W. Kleinlein, in Solid Oxide Fuel Cells V, U. Stimming,
S. C. Singhal, H. Tagawa, and W. Lehnert, Editors, PV 97-40, p. 746, The Elec-
trochemical Society Proceedings Series, Pennington, NJ ͑1997͒.
4. T. Horita, N. Sakai, and T. Kawada, Denki Kagaku oyobi Kogyo Butsuri Kagaku,
61, 760 ͑1993͒.
1/2
C x,t͒/C ϭ 1 Ϫ erf x/2 Dt͒
͔
͓2͔
͑
͓
͑
0
5. N. Lahl, K. Singh, L. Singheiser, K. Hilpert, and D. Bahadur, J. Mater. Sci., 38,
3089 ͑1999͒.
where C(x,t) is the concentration of ions measured at the distance,
x, from the surface of the ODS alloy for a given time period t. C0 is
the initial concentration of ions and D the diffusion coefficient. The
term erf denotes the error function and its values can be obtained
from the literature.10 The diffusion distance for the MASN6/ODS
couple was determined between the nominal Cr concentration in the
ODS alloy down to 1 atom % of Cr in the sealant. For the diffusion
couple MASC6/ODS, the distance was determined down to 2.5
6. N. Lahl, K. Singh, L. Singheiser, K. Hilpert, and D. Bahadur, in Solid Oxide Fuel
Cells ͑SOFC VI͒, S. C. Singhal and M. Dokiya, Editors, PV 99-19, p. 1057, The
Electrochemical Society Proceedings Series, Pennington, NJ ͑1999͒.
7. N. Lahl, K. Singh, L. Singheiser, K. Hilpert, and D. Bahadur, J. Electrochem. Soc.,
149, A607 ͑2002͒.
8. P. W. McMillan, Glass Ceramics, Academic Press, New York ͑1969͒.
9. J. H. Jean and T. K. Gupta, J. Mater. Res., 8, 2393 ͑1993͒.
10. P. G. Shewman, Diffusion in Solids, McGraw-Hill, New York ͑1970͒.
Downloaded on 2015-07-10 to IP 128.210.126.199 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).