Engineering Failure Analysis 29 (2013) 108–121

Contents lists available at SciVerse ScienceDirect

Engineering Failure Analysis

journal homepage: www.elsevier.com/locate/engfailanal

On the notch effect in load bearing capacity, apparent fracture

toughness and fracture mechanisms of polymer PMMA, aluminium alloy
Al7075-T651 and structural steels S275JR and S355J2
S. Cicero ⇑, V. Madrazo, T. García, J. Cuervo, E. Ruiz
Dpto. Ciencia e Ingeniería del Terreno y de los Materiales, Universidad de Cantabria, Av./Los. Castros s/n, 39005 Santander, Cantabria, Spain

a r t i c l e i n f o a b s t r a c t

Article history: This paper presents an analysis of the notch effect in the load bearing capacity, the appar-
Received 11 September 2012 ent fracture toughness and the fracture mechanisms of four different materials: polymer
Received in revised form 5 November 2012 PMMA, aluminium alloy Al7075-T651, and structural steels S275JR and S355J2 within their
Accepted 22 November 2012
corresponding ductile-to-brittle transition zone.
Available online 6 December 2012
Concerning the load bearing capacity and the apparent fracture toughness, a clear notch
effect has been observed. In the case of the apparent fracture toughness, this has been ade-
Keywords:
quately predicted here through the Theory of Critical Distances.
Notch effect
Load bearing capacity
As for fracture mechanisms, it is shown how these progressively change with the notch
Apparent fracture toughness radius, from basically brittle ones in cracked conditions, to non-linear mechanisms
Theory of Critical Distances observed for higher notch radii. This evolution has been observed in the four analysed
materials and, in all cases, has justified the corresponding consequences of the notch effect
on both the load bearing capacity and the apparent fracture toughness.
Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction

There are many situations where the defects or the stress risers that may cause a given structural failure are not neces-
sarily sharp. Some examples are defects such as mechanical damage, corrosion defects or fabrication defects, and structural
details, such as holes, corners and welded joints. Many of these are notch-type defects.
In this sense, components containing notches exhibit an apparent fracture toughness that is greater than that obtained in
cracked components because of the lower stresses acting at the notch tip for a given load (e.g., [1–7]). This generally has
direct consequences on the load bearing capacity of the component and also on the corresponding structural integrity assess-
ments [7,8].
At the same time, recent years have seen a great deal of research dedicated to providing an analytical notch theory capa-
ble of predicting the fracture behaviour of notched components, proposing different failure criteria [1,2]: the global fracture
criteria and the local fracture criteria. The latter have more practical applications, especially those based on the Theory of
Critical Distances (TCDs). The TCD is actually a group of methodologies, all of them using a characteristic material length
parameter (the critical distance) when performing fracture assessments (e.g., [9–12]). This length parameter is usually re-
ferred as L (critical distance), whose expression is (in fracture analysis):
 2
1 K IC
L¼ ð1Þ
p r0

⇑ Corresponding author.
E-mail address: [email protected] (S. Cicero).

1350-6307/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.engfailanal.2012.11.010
 S. Cicero et al. / Engineering Failure Analysis 29 (2013) 108–121 109

where KIC is the material fracture toughness and r0 is a characteristic material strength parameter (the inherent strength),
usually larger than the ultimate tensile strength (ru), that must be calibrated. Only in those situations where there is a lin-
ear-elastic behaviour at both the micro and the macroscale (e.g., fracture of ceramics) does r0 coincide with ru.
Within the TCD two methodologies stand out: the Point Method (PM) and the Line Method (LM), both based on the stress
field and providing very similar predictions [9]. The PM is the simplest methodology, and it assumes that fracture occurs
when the stress reaches the inherent strength (r0) at a distance of L/2 from the defect tip. The Line Method (LM), meanwhile,
assumes that fracture occurs when the average stress along a distance of 2L reaches the inherent strength, r0. Eqs. (2) and
(3), respectively, summarise both approaches:
 
L
r ¼ r0 ð2Þ
2

Z 2L
1
rðrÞdr ¼ r0 ð3Þ
2L 0

Therefore, the TCD allows the fracture assessment of components with any kind of stress riser to be performed. For exam-
ple, when using the PM in a given material, it would be sufficient to perform two fracture tests on two specimens with dif-
ferent types of defects (e.g., sharp notch and blunt notch). The specimens can then be modelled with finite elements,
determining the stress state at the notch tip at the corresponding fracture load and representing the resulting stress–dis-
tance curves. These curves cross each other at a point with coordinates (L/2, r0), as shown in Fig. 1, provided both are suf-
ficiently sharp. If it is intended to predict the fracture load of any other component (made of the same material) containing
any other kind of defect, both the component and the defect should be modelled, the fracture load being that one for which
Eq. (2) is fulfilled, L and r0 being those values obtained from the calibration process referred to above.
Moreover, both the PM and the LM provide expressions (whose justification may be found in [9,13]) for the apparent frac-
ture toughness (KIN) exhibited by notched components. The expressions derived from the PM [9] and the LM for U-shaped
notches when assuming the Creager–Paris stress distribution at the notch tip [14] are, respectively, in the following
equations:
 3=2
1þq
K IN ¼ K IC  L  ð4Þ
1 þ 2Lq

rffiffiffiffiffiffiffiffiffiffiffiffiffiffi
q
K IN ¼ K IC 1þ ð5Þ
4L

q being the notch radius.

These expressions may be used as another possibility for the calibration of the TCD parameters, consisting in performing
fracture tests on fracture specimens with different notch radii and fitting Eq. (4) (in case of using the PM) to the experimental
results, L being the fitting parameter. Thus, there are two basic ways of estimating L: by a combination of a limited number of
fracture tests and finite elements modelling, and by a number of fracture tests and numerical fitting.
Finally, in recent works (e.g., [15,16]), there has been additional validation of the TCD, also providing evidence about the
evolution of fracture micromechanisms when the notch radius changes. This paper presents a summary of this research, with
the addition of new experimental evidence obtained in structural steels, providing an overview of the evolution of fracture

Fig. 1. Obtaining the L and r0 parameters.

110 S. Cicero et al. / Engineering Failure Analysis 29 (2013) 108–121

micromechanisms in very different materials. Finally, it is analysed how this evolution has direct consequences on the above
mentioned load bearing capacity.

2. Experimental programmes

In order to analyse the notch effect covering a significant number of materials and situations, the following sets of spec-
imens are analysed:

– 32 Single Edge Notched Bending (SENB) specimens made of polymethyl methacrylate (PMMA) containing U-shaped
notches and notch radii varying from 0 mm (crack) to 2.5 mm. The thickness of the specimens was 5 mm and the
relation between the defect length and the specimen width (a/W) was 0.5 for all the notch specimens (and around
the same value for the cracked one). The main mechanical properties of the material are shown in Table 1. Details of
the experimental process, which was developed at room temperature, are gathered in [15], the geometry of the spec-
imens being shown in Fig. 2a

Table 1
Mechanical properties at room temperature of the four analysed materials. E: elastic modulus; ry: yield stress or 0.2% proof strength; ru: ultimate tensile
strength; emax: maximum strain.

E (GPa) ry (MPa) ru (MPa) emax (%)

PMMA 3.42 48.5 71.9 4.1
Al7075-T651 (longitudinal) 71.6 554.1 612.0 10.1
Al7075-T651 (transversal) 74.4 539.2 602.2 9.9
S275JR 203.0 328.0 518.5 15.9
S355J2 206.1 376.4 557.6 15.4

Fig. 2. Schematic showing the geometry of the specimens (dimensions in mm): (a) PMMA specimens, q varying from 0 mm to 2.5 mm; (b) Al7075-T651
specimens, q varying from 0 mm to 2.0 mm; (c) S275JR and S355J2 specimens, q varying from 0 mm to 2.0 mm.
 S. Cicero et al. / Engineering Failure Analysis 29 (2013) 108–121 111

Table 2
Load bearing capacity (LBC) obtained for the different combinations of material and notch radius. KIN and KJN in MPa m1/2.

Spec. PMMA Al7075-T651 S275JR S355J2

Notch radius (mm) LBC (N) KIN Notch radius (mm) LBC (kN) KIN Notch radius (mm) LBC (kN) KJN LBC (kN) KJN
1 0 130.03 2.31 0 11.78 28.75 0 24.08 61.32 21.04 44.33
2 83.00 1.62 11.32 26.90 38.2 88.05 30.47 63.34
3 131.23 2.19 10.50 25.82 34.61 78.17 34.48 74.15
4 0.25 124.90 2.66 12.30 28.17 34.90 95.03 –
5 119.90 2.55 9.81 25.71 0.15 59.93 283.2 71.70 143.2
6 104.00 2.21 10.46 24.50 58.56 246.3 62.50 118.0
7 107.10 2.28 0.15 20.95 50.59 65.25 392.5 58.51 110.9
8 0.32 117.40 2.50 21.31 51.46 64.22 379.9 31.97 54.66
9 112.60 2.40 18.95 45.77 0.25 57.00 223.7 65.47 126.8
10 102.50 2.18 20.03 48.38 57.87 260.9 78.34 175.8
11 108.70 2.31 19.65 47.47 57.11 246.6 60.29 115.1
12 0.50 90.00 1.92 21.31 51.47 53.63 170.0 58.52 110.4
13 85.20 1.81 0.21 23.68 57.18 0.50 69.25 481.9 81.00 220.2
14 170.30 3.63 22.67 54.76 61.59 330.1 77.45 341.7
15 162.60 3.46 22.79 55.03 61.18 321.8 78.71 256.9
16 1.0 212.80 4.53 0.47 31.71 76.57 69.86 501.9 76.74 179.0
17 213.60 4.55 30.13 72.77 1.0 59.59 302.5 82.45 407.4
18 204.80 4.36 32.33 78.08 66.38 437.7 82.18 541.7
19 202.80 4.32 1.0 39.70 95.88 63.55 374.0 89.89 350.5
20 202.60 4.31 39.17 94.59 72.54 575.2 82.42 266.4
21 1.5 215.50 4.59 37.94 91.63 2.0 80.85 950.3 97.01 688.7
22 165.90 3.53 2.0 44.57 107.64 80.26 976.2 87.83 448.0
23 219.00 4.66 44.95 108.56 79.84 947.4 80.86 208.8
24 197.90 4.21 45.23 109.23 78.95 897.2 –
25 2.0 258.50 5.51
26 261.10 5.56
27 237.80 5.06
28 2.5 253.80 5.41
29 259.90 5.54
30 250.40 5.33
31 251.30 5.35
32 243.20 5.18

Fig. 3. (a) Notch effect on the normalised load bearing capacity (LBC0 being the average LBC in cracked conditions); (b) notch effect on the apparent fracture
toughness, and predictions provided by the Point Method. PMMA.

– 24 Compact Tension (CT) specimens made of Al7975-T651 aluminium alloy, with U-shaped notches and notch radii
varying from 0 mm (crack) to 2.0 mm, and following TL orientation (see Fig. 2b). The thickness of the specimens was
20 mm, with the same a/W relation as that mentioned for the PMMA specimens in the case of notched specimens and
around this value in the case of cracked specimens. Table 1 gathers the corresponding mechanical properties. Details
of the experimental process, also developed at room temperature, may be found in [16].
– 24 CT specimens made of structural steel S275JR and containing U-shaped notches with six different notch radii vary-
ing from 0 mm to 2.0 mm (0, 0.15, 0.25, 0.50, 1.0 and 2.0 mm). The geometry of the specimens is shown in Fig. 2c.
Additional cracked specimens were tested at 10 °C following [17] in order to obtain the material transition temper-
112 S. Cicero et al. / Engineering Failure Analysis 29 (2013) 108–121

Fig. 4. (a) Notch effect on the normalised load bearing capacity (LBC0 being the average LBC in cracked conditions); (b) notch effect on the apparent fracture
toughness, and predictions provided by the Point Method. Al7075-T6751.

Fig. 5. (a) Notch effect on the normalised load bearing capacity (LBC0 being the average LBC in cracked conditions); (b) notch effect on the apparent fracture
toughness, and predictions provided by the Point Method. Structural steel S275JR, L = 0.004 mm.

Fig. 6. (a) Notch effect on the normalised load bearing capacity (LBC0 being the average LBC in cracked conditions); (b) notch effect on the apparent fracture
toughness, and predictions provided by the Point Method. Structural steel S355J2, L = 0.006 mm.

ature, T0 [18,19], providing a value of 26 °C. The 24 CT notched specimens were then tested at 50 °C, which cor-
responds to a temperature located within the lower part of the material ductile-to-brittle transition zone. The thick-
ness of the specimens was 25 mm, a/W being 0.5 for the notched specimens and close to this value for the cracked
ones. Table 1 gathers the corresponding mechanical properties at room temperature.
 S. Cicero et al. / Engineering Failure Analysis 29 (2013) 108–121 113

Fig. 7. Fracture micromechanisms observed in PMMA: (a) q = 0 mm, specimen 2; (b) 0.25 mm, specimen 7; (c) q = 0.32 mm, specimen 10; (d) q = 1.0 mm,
specimen 17; (e) q = 1.5 mm, specimen 24; (f) q = 2.0 mm, specimen 25. The white arrows indicate the initial notch front.

– 24 CT specimens made of structural steel S355J2 and containing U-shaped notches with six different notch radii vary-
ing from 0 mm to 2.0 mm (0, 0.15, 0.25, 0.50, 1.0 and 2.0 mm), as shown in Fig. 2c. Again, additional cracked spec-
imens were tested at 100 °C following [19] in order to obtain the material transition temperature, T0, this time
providing a value of 133 °C. The 24 CT notched specimens were then tested at 150 °C, which also corresponds
to a temperature located within the lower part of the material ductile-to-brittle transition zone. The thickness of
the specimens was 25 mm in all cases, a/W being 0.5 for the notched specimens and close to this value for the
cracked ones. Table 1 gathers the material mechanical properties at room temperature.
114 S. Cicero et al. / Engineering Failure Analysis 29 (2013) 108–121

Fig. 8. Fracture micromechanisms observed in PMMA specimens with q = 0.5 mm. (a) specimen 13, with brittle behaviour; (b) specimen 14, with non-
linear mechanisms.

Fig. 9. Non-linear fracture features and initiation point observed in PMMA notched specimen 24.

PMMA specimens were tested following ASTM D5045 [20], while both the Al7075-T 651 and the steel specimens were
tested following [17]. In order to achieve the required temperatures during the tests performed on S275JR and S355J2 CT
specimens, liquid nitrogen in combination with an insulating chamber was used.
Finally, it should be noted that the four materials have a basically brittle behaviour in cracked conditions, regardless of the
evolution in fracture micromechanisms that may take place when introducing a finite radius in the defect tip.

3. Results

3.1. Load bearing capacity and apparent fracture toughness

Table 2 shows the results, in terms of both the load bearing capacity (LBC) and the apparent fracture toughness, for the
four materials analysed. Figs. 3–6 also gather the load bearing capacity results, this time graphically and normalised by the
average load bearing capacity in cracked conditions, together with the apparent fracture toughness results: KIN for PMMA
and Al7075-T651, and KJN for the two structural steels. The corresponding apparent fracture toughness for the different
materials and notch radii has been obtained [9] by the application of cracked specimen formulation [20] to the notched spec-
imens. Therefore, Eq. (6) has been used for the PMMA SENB specimens, Eq. (7) for the CT Al7075-T651 specimens, and Eqs.
(8) and (9) for the CT S275JR and S355J2 specimens.
 
0         2 1
P max  a 0:5 1:99  Wa  1  Wa  2:15  3:93 Wa þ 2:7 Wa
K IN ¼ 6 @    1:5 A ð6Þ
B  W 0:5 W 1þ2 a  1 aW W
 S. Cicero et al. / Engineering Failure Analysis 29 (2013) 108–121 115

(a) (b)

Precracked surface

(c) (d)

(e) (f)

Fig. 10. Fracture micromechanisms observed in Al7075-T651 (TL orientation): (a) q = 0 mm, specimen 2; (b) 0.15 mm, specimen 12; (c) q = 0.21 mm,
specimen 15; (d) q = 0.47 mm, specimen 17; (e) q = 1.0 mm, specimen 21; (f) q = 2.0 mm, specimen 23. The white arrows indicate the initial notch front.

   a  a 2  a 3  a 4 
Pmax 2 þ Wa
K IN ¼    0:886 þ 4:64  13:32 þ 14:72  5:60 ð7Þ
B  W 0:5 1  Wa
3=2 W W W W

where Pmax is the corresponding maximum load (Table 2), B is the specimen thickness, W is the specimen width, and a is the
defect length [17,20].
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
E
K JN ¼ J CN ð8Þ
1  t2
where JCN is the apparent J-integral at onset of cleavage fracture, E is the Young’s modulus and t is the Poisson’s ratio [20].
116 S. Cicero et al. / Engineering Failure Analysis 29 (2013) 108–121

Fig. 11. Detail of the flat area observed in Al7075-T651 specimen 21.

ð1  t2 ÞK 2eN gAp
J CN ¼ J eN þ J pN ¼ þ ð9Þ
E Bb0
where JeN and JpN are, respectively, the elastic and plastic components of JCN, KeN is the apparent elastic stress intensity factor
at onset of cleavage, g is a dimensionless constant, Ap is the plastic area under the load–displacement curve and b0 is the
initial remaining ligament [20]. Here, it should be noted that when the KJN is being calculated, it does not only depend on
the LBC determined in the corresponding load–displacement curve, but also on Ap. This is the reason why specimens with
similar LBC may have very different KJN (e.g., 355J2 specimens with a notch radius of 2.0 mm).
The apparent fracture toughness has also been adjusted for the four materials by using the PM predictions (Eq. (4)). The
values of L for PMMA and Al7075 have been taken from [15,16], respectively, whereas the L values for S275JR and S355J2
steels have been obtained from the best fit curve. The KIC values (KJC in the case of the two steels), used in Eq. (4), have been
taken as the average of the results obtained in cracked conditions (2.04 MPa m1/2 for the PMMA, 26.65 MPa m1/2 for the
Al7075-T651, 80.64 MPa m1/2 for the S275JR steel, and 60.60 MPa m1/2 for the S355J2 steel).
The main observations derived from the graphs showing the evolution of the normalised LBC are the following (Figs. 3–6a):

– The four materials present a clear notch effect on the LBC, although Al7075-T651 presents the largest one, with a LBC
almost four times that observed in cracked conditions (LBC0) for a notch radius of 2.0 mm. Steel S355J2 presents a 3.4
times larger LBC for the same notch radius, whereas this factor is 2.4 for steel S275JR and 2.2 for the PMMA.
– The evolution of the LBC presents singularities for the different materials. Firstly, the PMMA has very little notch
effect for notch radii below 0.5 mm, whereas the other three materials present significant increases in the LBC even
for the smallest radii being considered.
– Secondly, Al7075-T651 presents little scatter when compared to the PMMA and the two structural steels.
– Thirdly, both the S275JR and the S355J2 steels present an abrupt notch effect for the smaller notch radii, after which
the notch effect increases very little (at 0.15 mm for the S275JR steel and at 0.25 mm for the S355J2 steel).
– In contrast, the notch effect on PMMA and Al7075-T651 increases significantly along a larger range of notch radii (up
to a notch radii of 2.0 mm could be taken as a reference).

The results obtained in terms of apparent fracture toughness (against the square root of the notch radius) also provide
relevant observations:

– The curves present a much more continuous increase than the LBC curves. The notch effect on this parameter is not
decelerated, as occurs with the LBC.
– The two steels present the higher notch effect when considering KJN. Average values for the 2.0 mm notch radius are
between 10.7 (S275JR) and 8.2 (S355J2) times larger than the average values observed in cracked conditions.
– Small variations in the LBC may result in very significant variations in KJN, because this latter parameter also depends
on Ap, and not only on the LBC. Concerning KIN, this cannot be the case, as this parameter is directly proportional to
the LBC (Pmax in Eqs. (6) and (7)).
 S. Cicero et al. / Engineering Failure Analysis 29 (2013) 108–121 117

Fig. 12. Fracture micromechanisms observed in steel S275JR (50 °C): (a) q = 0 mm, specimen 3; (b) 0.15 mm, specimen 5; (c) q = 0.25 mm, specimen 9; (d)
q = 0.50 mm, specimen 13; (e) q = 1.0 mm, specimen 18; (f) q = 2.0 mm, specimen 23. The white arrows indicate the initial notch front.

– As happened with the LBC, and exclusively for the PMMA specimens, there is a notch radius (critical radius [9,15])
below which there is no notch effect, or where it is not significant. This is not the case for the other three materials.
This phenomenon, with materials having such a critical radius, and materials where this phenomenon is not
observed, has been previously reported in the literature (e.g., [9])
– The scatter in the results obtained for notched specimens within the ductile-to-brittle transition zone of the two
structural steels is quite significant, even larger than that observed in cracked conditions.
118 S. Cicero et al. / Engineering Failure Analysis 29 (2013) 108–121

Fig. 13. (a) Brittle fracture (cleavages) observed in S275JR specimen 3 (q = 0 mm); (b) ductile fracture (microvoids) found after the initial crack front in
S275JR specimen 5 (q = 0.15 mm).

3.2. Analysis of SEM fractographies

This section analyses the notch effect on the fracture mechanisms of the four materials being analysed. In all cases, from
Figs. 7–14, the specimens that have been selected for the analysis are those presenting the KIN or KJN result that is closest to
the corresponding average value (for each notch radius).
Fig. 7 gathers examples, one per notch radius, of the fracture surface around the initial notch front in PMMA specimens.
This allows some general observations to be made on the notch effect. It can be observed that the specimens with lower
notch radii (0 mm, 0.25 mm and 0.32 mm) have very similar fracture features. The fracture has a brittle aspect, and corre-
sponds to the mirror zone [21–24] (a zone where thin planar crazes form a flat smooth fracture origin associated to slow
crack growth). It is also possible to identify ‘‘initiation lines’’, which are actually interference colour fringes caused by the
presence of a thin layer of highly oriented polymer (crazing) with a different refractive index from that of the bulk. It can
be observed how the size of these lines increases with the notch radius [22]. Therefore, specimens with lower notch radii,
including cracks, have very similar fracture mechanisms, leading to very similar results in terms of LBC and KIN.
Fig. 7d–f shows the fracture surfaces for the larger radii (1.0 mm, 1.5 mm and 2.0 mm). This time there are ‘‘initiation
areas’’, instead of ‘‘initiation lines’’, whose size increases with the notch radius and are related to the development of
non-linear processes [15]. For the sake of simplicity, the results for 2.5 notch radii are not shown, as they are basically
the same as those shown in Fig. 7d–f. These similar micromechanisms have led to similar LBC and KIN results.
The fractographies obtained for specimens with 0.5 notch radius have not been included deliberately in Fig. 7, as they
present two very distinct cases, and are shown in Fig. 8. Two of them, those corresponding to specimens 12 and 13, have
very similar fracture micromechanisms to those found in specimens with lower radii (brittle aspect), whereas the other
two (specimens 14 and 15) present non-linear mechanisms which are very similar to those observed in higher radii. This
has direct consequences on both the LBC and KIN: specimens 12 and 13 provide low values of these variables, very similar
to those observed in cracks, without any notch effect; specimens 14 and 15 present a clear increase (notch effect) in these
two variables, because of the development of the non-linear processes appearing in larger radii. This notch radius, 0.5 mm, is
then a transition radius on which two very different fracture behaviours may be observed.
Finally, Fig. 9 presents a detail of the fracture mechanisms observed in the initiation areas, with clear non-linear features
and an initiation point.
Concerning the aluminium alloy, Fig. 10 presents some of the observed fracture surfaces. In the case of Fig. 10a, corre-
sponding to cracked specimen 2, a direct transition can be observed from the precracked surface to a brittle fracture surface,
and also that the fracture features have a certain orientation in the same direction as the crack propagation (perpendicular to
the initial crack front). The fracture seems to be transgranular.
Fig. 10b shows an image of Al7075-T651 specimen 12, with a notch radius of 0.15 mm. It can be observed that there is a
flat initiation area, from 50 lm to 100 lm thick, between the initial crack front and the brittle fracture surfaces, something
not observed in any of the cracked specimens. The aspect of these areas is smoother than the rest of the fracture surface.
There is also a certain orientation of the brittle fracture features in the crack propagation direction. After the flat areas, in
the propagation direction, fracture is transgranular.
Fig. 10c–f, corresponding to specimens Al7075-T651 specimens 15, 17, 21 and 23, present flat initiation areas which are
distributed all along the initial crack front. These flat areas increase their size with the notch radius, and are probably caused
by the reduction of triaxiality on the defect tip, which allows non-linear processes to develop within this zone. Fig. 11 shows
a detail of flat area observed in specimen 21, revealing ductile fracture mechanisms with shallow dimples.
 S. Cicero et al. / Engineering Failure Analysis 29 (2013) 108–121 119

Fig. 14. Fracture micromechanisms observed in steel S355J2 (150 °C): (a) q = 0 mm, specimen 2; (b) 0.15 mm, specimen 7; (c) q = 0.25 mm, specimen 9;
(d) q = 0.50 mm, specimen 15; (e) q = 1.0 mm, specimen 17; (f) q = 2.0 mm, specimen 22. The white arrows indicate the initial notch front.

Analogously to what happened in PMMA, generally, the larger the notch radii, the larger the size of the flat initiation areas
and the more ductile the fracture is. However, in the case of Al7075-T651, the flat areas increase their size all along the notch
range considered, and therefore there is a continuous evolution of both the LBC and the KIN within this range. Hence, again,
the notch effect corresponds to an evolution in the fracture micromechanisms, from brittle ones to others in which the ini-
tiation of fracture is preceded by the development of initiation areas associated to more ductile mechanisms; the larger the
notch radius, the larger the size of such initiation areas.
120 S. Cicero et al. / Engineering Failure Analysis 29 (2013) 108–121

Fig. 15. Fracture micromechanisms observed in S355J2 specimens with 2.0 mm notch radius (150 °C): (a) specimen 21 (maximum KJN); (b) specimen 23
(minimum KJN). The white arrows indicate the initial notch front.

Regarding structural steel S275JR, Fig. 12 presents the corresponding evolution of the fracture micromechanisms. The
observations are analogous to those seen in the aluminium alloy. Firstly, the cracked specimen has a basically direct tran-
sition from the precracking surface to the final fracture surface, where multiple cleavages can be observed (see Fig. 13). There
is a very narrow whitish line along the initial crack front, in the middle of Fig. 12a, that may be an indication of non-linear
processes, but it is of very limited thickness (around 10 lm). In any case, cleavage totally dominates the fracture process.
Secondly, once a finite notch radius is introduced in the specimen, non-linear processes appear. Fig. 12b shows an image
of S275JR specimen 5, corresponding to a notch radius of 0.15 mm. It can be seen how there is an initiation area after the
initial crack that does not correspond to cleavage fracture. Fig. 13b shows a detail of this area, revealing the presence of
microvoids. Finally, after the initiation area, brittle fracture dominates again, with analogous mechanisms to those shown
in Fig. 13a.
Finally, the size of the initiation areas generally grows with the notch radius, something that has a direct effect on the KJN
results. The effect on the LBC is more moderate, once the 0.15 mm notch radius has been exceeded.
To finish with the SEM analysis, Fig. 14 gathers the evolution of the fracture micromechanisms in steel S355J2. The obser-
vations are quite similar to those seen in both the aluminium alloy and steel S275JR, but here, although there is a clear evo-
lution of fracture micromechasims when observing the whole sequence of fractographies, there is no direct link between
them and the LBC and KIN measurements: whereas the latter increases significantly for very small notch radii (e.g.,
0.15 mm and 0.25), the change in fracture micromechanisms appear for larger radii (1.0 mm). Thus, the increase in the mac-
roscopic measurements takes place first, and afterwards, the change in the fracture micromechanisms.
Another important fact is the high scatter observed in the KJN results of both steels. It is well known that KJC results ob-
tained in cracked conditions within the ductile-to-brittle transition zone of ferritic–pearlitic steels may present a high degree
of scatter, but here it has been observed that this scatter is even more significant in notched conditions. The results observed
in steel S355J2 with a notch radius of 2.0 mm are the most significant in this sense, where the maximum value (688.7 MPa m1/
2
) is 3.3 times larger than the minimum one (208.8 MPa m1/2). Again, this has a clear explanation when the corresponding
fracture surfaces are analysed. Fig. 15 presents the fracture surfaces of steel S355J2 specimens 21 (maximum KJN value)
and 23 (minimum KJN value). The former (Fig. 15a) presents a continuous initiation area containing ductile mechanisms (anal-
ogous to those shown in Fig. 13b), whereas the latter (Fig. 15b) presents cleavages as the dominant fracture mechanism. These
different fracture observations have led to significant differences in the LBC (around 20%), and huge differences in KJN (around
330%). The different effect on KJN when compared to the effect on LBC has been explained above (Section 3.1). Specimen 22,
with intermediate results in terms of both the LBC and KJN, also presents an intermediate situation concerning fracture mech-
anisms, showing an initiation area with non-linear mechanisms whose size is smaller to that seen in specimen 21.

4. Conclusions

This paper has presented the results and the corresponding study of different experimental programs performed in four
different materials (PMMM, Al7075-T651, S275JR and S355J2) with the aim of analysing the notch effect on the load bearing
capacity (LBC), the apparent fracture toughness (KIN or KJN) and fracture micromechanisms. The relation between these dif-
ferent variables has also been analysed.
It can be concluded that there is a clear notch effect on the four materials, although there are significant particularities on
each case:
 S. Cicero et al. / Engineering Failure Analysis 29 (2013) 108–121 121

(1) PMMA presents a critical radius below which there is no notch effect, something that has not been observed in the
other three materials.
(2) Al7075-T651 has presented the higher notch effect on the LBC (with an increase of almost 400% for a 2.0 mm notch
radius). This material has presented the lowest scatter in the results for the different notch radius, and has also pre-
sented a continuous increase in both the LBC and the KIN results within the range of notch radii considered here.
(3) The two ferritic–pearlitic steels have presented the highest notch effect on the apparent fracture toughness.
(4) Most of the notch effect in terms of LBC observed in the two steels is immediate (i.e., it takes place when introducing
the smallest notch radius, 0.15 mm), while this effect is much more moderate once the notch radius continues its
growth. However, the corresponding effect on KJN has a continuous significant growth, given that in spite of the small
increase in the LBC, the behaviour becomes more and more non-linear, with a direct impact on the plastic component
of KJN.
(5) The well-known scatter associated to KJC results observed in cracked specimens within the ductile-to-brittle transition
zone of ferritic–pearlitic steels does also take place in notched conditions (then, in KJN). Moreover, here, the scatter
observed in the notched specimens has been greater than that observed in the cracked specimens.
(6) The capacity of the Theory of Critical Distances (its Point Method version) for predicting the evolution of the apparent
fracture toughness has been validated.
(7) A relation has been established between the macroscopic observations (LBC, KIN, KJN) and the fracture micromecha-
nisms. This relation explains not only the general evolution of the macroscopic variables, but also the scatter observed
in the results.

Acknowledgements

The authors of this work would like to express their gratitude to the Spanish Ministry of Science and Innovation for the
financial support of the project MAT2010-15721: ‘Análisis de integridad estructural en defectos tipo entalla’, on the results of
which this paper is based.

References

[1] Niu LS, Chehimi C, Pluvinage G. Stress field near a large blunted V notch and application of the concept of notch stress intensity factor to the fracture of
very brittle materials. Eng Fract Mech 1994;49:325–35.
[2] Pluvinage G. Fatigue and fracture emanating from notch; the use of the notch stress intensity factor. Nucl Eng Des 1998;185:173–84.
[3] Bao Y, Jin Z. Size effects and mean strength criterion for ceramics. Fatigue Fract Eng Mater Struct 1993;16:829–35.
[4] Fenghui W. Prediction of intrinsic fracture toughness for brittle materials from the apparent toughness of notched-crack specimen. J Mater Sci
2000;35:2543–6.
[5] Seweryn A, Lukaszewicz A. Verification of brittle fracture criteria for elements with V-shaped notches. Eng Fract Mech 2002;69:1487–510.
[6] Cicero S, Gutiérrez-Solana F, Álvarez JA. Structural integrity assessment of components subjected to low constraint conditions. Eng Fract Mech
2008;75:3038–59.
[7] Cicero S, Madrazo V, Carrascal IA, Cicero R. Assessment of notched structural components using failure assessment diagrams and the Theory Of Critical
Distances. Eng Fract Mech 2011;78:2809–25.
[8] Horn AJ, Sherry AH. An engineering assessment methodology for non-sharp defects in steel structures – Part I: Procedure development. Int J Press Vess
Pip 2010;87:137–50.
[9] Taylor D. The theory of critical distances: a new perspective in fracture mechanics. Elsevier; 2007.
[10] Taylor D, Merlo M, Pegley R, Cavatorta MP. The effect of stress concentrations on the fracture strength of polymethylmethacrylate. Mater Sci Eng
2004;A382:288–94.
[11] Neuber H. Theory of notch stresses: principles for exact calculation of strength with reference to structural form and material. Berlin: Springer Verlag;
1958.
[12] Peterson RE. Notch sensitivity. In: Sines G, Waisman JL, editors. Metal fatigue. New York: McGraw Hill; 1959. p. 293–306.
[13] Taylor D, Cornetti P, Pugno N. The fracture mechanics of finite crack extension. Eng Fract Mech 2005;72:1021–38.
[14] Creager M, Paris C. Elastic field equations for blunt cracks with reference to stress corrosion cracking. Int J Fract 1967;3:247–52.
[15] Cicero S, Madrazo V, Carrascal IA. Analysis of notch effect in PMMA by using the theory of critical distances. Eng Fract Mech 2012;86:56–72.
[16] Madrazo V, Cicero S, Carrascal IA. On the point method and the line method notch effect predictions in Al7075-T651. Eng Fract Mech 2012;79:363–79.
[17] ASTM E1820-09e1. Standard test method for measurement of fracture toughness. Philadelphia: American Society for Testing and Materials; 2009.
[18] Wallin K. The scatter in KIC results. Eng Fract Mech 1984;19:1085–93.
[19] ASTM E1921-05. Standard test method for determination of reference temperature, T0, for ferritic steels in the transition range. Philadelphia: American
Society for Testing and Materials; 2005.
[20] ASTM D5045-99. Standard test methods for plane-strain fracture toughness and strain energy release rate of plastic materials. Philadelphia: American
Society of Testing and Materials; 1999.
[21] Kusy RP, Turner DJ. Influence of molecular weight of poly(methyl methacrylate) on fracture morphology in notched tension. Polymer
1977;18:391–402.
[22] So PK. Fractography, engineered materials handbook. vol. 2, Engineering plastics. Metals Park (OH): ASM International; 1988, p. 805–10.
[23] Liu K, Piggcyit MR. Fracture failure processes in polymers. I: Mechanical tests and results. Polym Eng Sci 1998;38:60–8.
[24] Liu K, Piggcyit MR. Fracture failure processes in polymers. II: Fractographic evidence. Polym Eng Sci 1998;38:69–78.