Mechanical Characterization of Solid Oxide Fuel Cells and Sealants

Figure
5-3
) show a significantly longer plateau 
region in case of specimens which were previously loaded with a slow displacement rate to 
obtain a pre-crack before the DT test. The specimens tested without previously inducing a pre-
crack showed a shorter and a slightly instable plateau region (
Figure
 5-3
). However, when taking 
individual thickness of specimen and average value of the plateau region into account, fracture 
toughness was identical within the limits of uncertainty. The average fracture toughness of 
specimens with and without the pre-crack was 3.05 ± 0.17 MPa·m
1/2
(
Table
 5-1
), confirming that 
there is no necessity to pre-crack the current thin and brittle material.
The obtained fracture toughness values are slightly higher than the ones reported in literature for 
NiO-3YSZ (2.55 ± 0.2 MPa·m
1/2 
[138]), which might be a result of slight porosity differences of 
the respective specimen batches. Furthermore, the values are significantly lower than reported 
for dense 3YSZ, which has a fracture toughness of 5.0 MPa·m
1/2
[195]. It might be assumed that 
the difference in fracture toughness is mainly caused by a difference in porosity of the specimens. 
1
Study has partly been published as J. Wei, G. Pećanac, et al., Proc. 12th Euro SOFC & SOE Forum, Luzern, 
B1309, 2016 and publiched as G. Pećanac, J. Wei, J. Malzbender, J Power Sources, 327 (2016) 629-637. 

---

Results and discussion 
77 
Furthermore, Delaforce et al [196] showed that NiO also effects the stability of the t-phase, 
leading to less transformation toughening of 3YSZ and hence a lower fracture toughness.
Figure 5-3: Typical curves of fracture tests with and without a pre-crack, indicating similar 
results and therefore confirming that pre-crack is not necessary for the material in the used 
geometry. 
b) NiO-8YSZ 
Similar initial tests were carried out on NiO-8YSZ (type 4) materials to analyze the necessity of 
pre-cracking. The NiO-8YSZ specimens are slightly thicker, hence, the elastic energy before 
crack initiation would be much larger and the instable crack growth might be possible in this 
case. The resulting curves confirmed this assumption crack failed instable and did not permit a reliable fracture toughness characterization (red curve 
Compared to the non-pre-cracked specimen, the pre-cracked specimen yielded a stable crack 
growth region, which permitted a reliable determination of fracture toughness. The obtained 
average fracture toughness of 1.76 ± 0.15 MPa·m
1/2
, considering the difference in material’s 
production and porosity, is in agreement with the reported data from Radovic et al ( 1.6 ± 0.2 for 
21.8 ± 1.3 % porosity [137]). According to literature, the fracture toughness of dense 8YSZ and 
NiO are 1.65 MPa·m
1/2
[197]
and 1.5 MPa·m
1/2
[198] respectively, implying that the difference 
compared to those values might be a result of experimental uncertainties, or it might be 

---

Results and discussion 
78 
speculated that NiO has even a positive influence on the fracture resistance of the composite. The 
latter might not be the case since Delaforce et al [196] found that in case of 8YSZ the phase 
composition remained unaltered after NiO was added. 
Figure 5-4: The type 4 Ni-8YSZ specimens were tested with and without pre-cracking. It showed 
that the fracture test without a pre-crack test leads instable crack growth and overestimation of