Mechanical Characterization of Solid Oxide Fuel Cells and Sealants

partner CISC, Madrid, the testing temperature was reduced to 650°C.
Crystallization is expected to be almost completed after 800 h heating treatment, which is 
confirmed by the similar fracture stress at room temperature and 800°C for annealed specimens.
A previous work [65] revealed an average fracture stress of 91 ± 12 MPa at room temperature 
and ~ 90 MPa at 800°C for Jülich bar-shaped B type sealant, which is a material sintered at 
950°C for 10 h. Compared with 7.5 B(Ba) and 10 B(Sr), sealant B has hence a superior fracture 
stability, while Jülich bar-shaped H-P sealant yielded much lower fracture stress 28 ± 3 MPa at 
room temperature and ~1 MPa at 800 °C, indicating a lower elevated temperature stability 
compared with 7.5 B(Ba) and 10 B(Sr).
Table 5-12: Average fracture stress (in MPa) obtained from 3-point-bending tests on bar type 
specimens. 
Sealant 
Pre-treatment 
RT 
650°C 
800°C 
7.5 B(Ba) 
24h/800°C 
81 ± 5 
25 ± 4 
800h/800°C 
63 ± 7 
72 ± 2 
10 B(Sr) 
24h/750°C 
101 ± 18 
92 ± 8 
800h/750°C 
68 ± 9 
84 ± 17 
5.2.5.
Shear strength
3
5.2.5.1.
Sealant H-Ag 
Torsion tests of the H-Ag sealant at RT after 10 h joining at 850 °C yielded similar an average 
fracture stress (61 ± 4 MPa) as obtained from bending tests on head-to-head specimens (55 ± 6 
MPa). A comparison of bending and shear stress of H-Ag sealant is shown in 
Figure
 5-38
. Two 
tests at 600°C resulted in a similar shear stress of 64.1 ± 0.3 MPa, indicating that residual 
3
Study has partly been submitted for publication as J. Wei, G. Pećanac, et al., Proc. 12th Euro. SOFC & SOE Forum, 
Luzern, B1309, 2016. 

---

Results and discussion 
122 
stresses induced by thermal mismatch of sealant and steel, which are maximum at RT, do not 
affect the torsion test. A single test at 800°C revealed a non-linear loading behavior of H-Ag 
specimens and a significant decrease in shear fracture stress (13 MPa), similar as previously 
observed for head-to-head joined bending specimen (7 MPa, see above). This decrease in the 
fracture stress can be associated with the softening of the sealant above the glass transition 
temperature 
T
g
due to residual glassy phases that also led to noticeable creep effects reported for 
head-to-head joined bending specimens.
Testing of a specimen after joining for 100 h at 850°C, to simulate the effect of longer stack 
operation, yielded a lower RT shear failure stress (14 MPa). A similar effect has been reported 
for annealed H-Ag head-to-head joined bending specimens. Such a decrease in fracture stress has 
obvious implications for the sealant application in stacks.
Failure in bending tests appeared to be predominant via cracking through the glass-ceramic close 
to the interface with the steel (oxide scale did not delaminate), i.e. the fracture path in case of RT 
and elevated temperature torsion test specimens was the interface of the oxide scale with the 
steel as shown in 
Figure
 5-39
.  
Figure 5-38: Comparison of bending and shear stress of H-Ag sealant.

---

Results and discussion 
123 
Figure 5-39: Fractured specimen of H-Ag after the torsional shear test at RT. 
5.2.5.2.
Sealant H-F 
Due to the load limit of the set-up it was not possible to induce RT failure for 10 h joined 
specimens, implying that the shear stress is higher than 68.6 MPa, indicating a much higher shear 
stress value than bending stress (13 ± 1 MPa) obtained from head-to-head specimens. At 800°C, 
a similar value (17 MPa) as for H-Ag was obtained, where the bending stress of H-F could not be 
determined due to extreme deformation of head-to-head specimen. The shear stresses obtained 
from torsion tests are listed in 
Table
 5-13
. 
For 100 h joined specimens the average torsional shear failure stress was 58 ± 8 MPa (for the 
single test where failure did not occur up to the maximum applicable stress of 68.6 MPa, this 
particular stress was assumed to be the failure stress), hence being slightly lower than for the 10 
h joining case. However, compared to bending stress (23 ± 1 MPa, see above) obtained from 
head-to-head specimens, the shear stress is still much higher (see 
Figure
5-40
). The value at 
600 °C was similar as at RT indicating again that residual stresses do not affect failure stress.
In an initial elevated temperature loading experiment at 660°C failure could not be initiated and 
deformation was dominated by plastic deformation of the steel. Nevertheless, the same specimen 
was subsequently loaded at 760°C, where failure of the sealant occurred (see 
Figure