Experiment 3: Fracture Toughness Testing of Aluminum

Performed February 19th, 2024.

Due March 11th, 2024.

MECH 321 TI-X

Group B

OBJECTIVE

The objective of the experiment is to measure the fracture toughness of various sized and
fatigued specimens to determine that they meet the required stress intensity factor.
INTRODUCTION

All materials always have flaws, as no material is perfect. In the engineering sense, it is essential
to avoid these flaws, thus it is necessary to understand the strength and behavior of cracks in
materials to provide reliable and consistent products composed of thousands of parts, that could
independently have different flaws. The stress around a crack is measured using the following:
𝐾
𝜎𝑦 =
√2𝜋𝑥
K = Stress intensity

Materials could be fractured or fatigued, as well as varying behavior under varying thicknesses,
this is tested by deforming the specimen and measuring the force required to do so, allowing to
find a connection between deformation and force applied. This property is called Fracture
toughness, 𝐾𝐼𝐶 , in this particular test, a uniaxial machine provides stress along an axis in tension,
this apparatus allows the measurement of displacement and force required through 2 different
sensors, the first is a DC-LVDT which measures the displacement of the crosshead, and the
second is a C.O.D gauge which measures the displacement of two accurately located knife edges.
In this experiment, due to the lack of variables and the controlled environment, the results are
expected to be relatively accurate, and a low deviation from previous models is expected.

Figure 1: Instron Tensile Machine [1]

To calculate the stress intensity at fracture, the equation below is used:
𝑃𝑄 2+𝛼
𝐾𝑄 = ( ) (0.866 + 4.64𝛼 − 13.32𝛼 2 + 14.72𝛼 3 − 5.6𝛼 4 ) ∗ [ ]
𝑡 √𝑤 (1 − 𝛼)1.5
ai = Total Crack length (cm)
t = Thickness (cm)
w = Width (cm)
PQ = Load (kN)
𝑎𝑖
α=
w

When the data is accumulated from the test in the software, load, and displacement curves are
made to evaluate the type of behaviour it exhibits to calculate the fracture toughness of the
specimen.

Figure 2: Three types of load-displacement behavior and identification of critical load. Type I –
tearing, Type II – mixed, Type III – cleavage [2].

Figure 2 describes the different types of relationships. Slope 𝑚𝑡 uses the elastic region to create a
linear function to go through the load and displacement curve. The 𝑚5 slope is a 5% offset from
𝑚𝑡 and helps determine the exact type of behaviour.
PROCEDURE

Before testing the specimen's fracture properties, their half height, width to hole centers, and
thickness were measured. The yield strength was already known for each specimen so that
property was also written down. Once the measurements for each specimen were taken, one at a
time, the specimens were installed into the Instron Tensile Machine and had their induced load
and displacement measured until they were fractured. Once the fractured specimen was removed
from the machine, its crack length was measured and recorded. The data from each specimen in
the software was transferred to a USB drive to calculate the remaining properties for each
specimen.

RESULTS

Table 1: Dimensional and Strength Properties of the four different specimens.

1.0 inch Non- 1.5-inch non- 1.5 inch 2.0 inch

Fatigued fatigued Fatigued Fatigued

Yield strength, MPa 276 276 276 276

h, mm 31.52 63.86 31.94 31.79

w, mm 50.225 45.1 56.99 50.72

t, mm 25.4 38.25 38.23 50.88

a_i, mm 22.86 22.86 26.17 28.46

 Elastic Slope of 1 inch Non Fatigued
45000

40000 y = 18564x - 5345.1

35000

30000
Force (N)

25000

Series1
20000
Linear (Series1)
15000

10000

5000

0
0 0.5 1 1.5 2 2.5 3
Displacement (mm)

Figure 3: Elastic slope of one inch non fatigued specimen.

 Force v Displacement 1 inch Non Fatigued
60000

50000

40000
Force (N)

30000 Series1
m_t

20000 m_s

10000

0
0 2 4 6 8 10 12 14
Displacement (mm)

Figure 4: Force v Displacement of one inch non fatigued specimen.

Figure 5: Fractured surface of 1.0-inch non-fatigued specimen.

Figure 6: Fracture Crack for 1.0-inch non-fatigued specimen.

 Elastic Slope of 1.5 inch Non Fatigued
70000

60000
y = 27497x - 9437.7

50000

40000
Force (N)

Series1
30000
Linear (Series1)

20000

10000

0
0 0.5 1 1.5 2 2.5 3
Displacement

Figure 7: Elastic slope of the 1.5-inch non-fatigued specimen.

 Force v Displacement 1.5 inch Non Fatigued
90000

80000

70000

60000
Force (N)

50000
Series1
40000 m_t
m_5
30000

20000

10000

0
0 1 2 3 4 5 6 7 8
Displacement (mm)

Figure 8: Force v Displacement of 1.5-inch non fatigued specimen.

Figure 9: Fractured surface of the 1.5-inch non-fatigued specimen.

Figure 10: Fracture crack for 1.5-inch non-fatigued specimen.

 Elastic Slope of 1.5 inch Fatigued
45000
y = 25030x - 6870.1
40000

35000

30000
Force (N)

25000

20000 Series1
Linear (Series1)
15000

10000

5000

0
0 0.5 1 1.5 2 2.5
Displacement (mm)

Figure 11: Elastic slope of the 1.5-inch fatigued specimen.

 Force v Displacement
60000

50000

40000
Force (N)

30000 Series1
m_t

20000 m_5

10000

0
0 2 4 6 8 10 12
Displacement (mm)

Figure 12: Force v Displacement of the 1.5-inch fatigued specimen.

Figure 13: Fractured surface of 1.5-inch Fatigued specimen.

Figure 14: Fracture crack of 1.5-inch Fatigued specimen.

 Elastic Slope of 2 inch Fatigued
35000

30000
y = 22158x - 6909.8

25000

20000
Force (N)

Series1
15000
Linear (Series1)

10000

5000

0
0 0.5 1 1.5 2
Displacement (mm)

Figure 15: Elastic slope of two-inch fatigued specimen.

 Force v Displacement of 2 inch Fatigued
50000

45000

40000

35000

30000
Force (N)

25000 Series1
m_t
20000
m_5
15000

10000

5000

0
0 1 2 3 4 5 6 7 8 9 10
Displacement (mm)

Figure 16: Force v Displacement of the two-inch fatigued specimen.

Figure 17: Fractured surface of 2.0-inch Fatigued specimen.

Figure 18: Fracture crack of 2.0-inch Fatigued specimen.

 Table 2: Fracture properties of different specimens.
1-inch non- 1.5-inch non- 1.5-inch 2-inch
fatigued fatigued Fatigued Fatigued
Determining load at fracture
The initial slope of elastic deformation
18546 27497 25030 22158
line mt
m5 = 95% mt 17635.8 26122.15 23778.5 21050.1
Load P5 [N] 39684.8 61384.7 40427.5 31577.7
Max Load Pmax [N] 41464.96 63410.6 41412.91 32130.7
𝑃𝑄 [𝑁] 39684.8 63410.6 40427.5 32130.7

Determining Stress Intensity at Fracture

α= ai/w 0.455151817 0.506874 0.459203 0.56112

𝐾𝑄 [𝑀𝑃𝑎 √𝑚] 58.141 75.895 37.370 32.633

Determining valid KIC

2.5(KQ/σ0)2 [m] 0.1109 0.18904 0.0445832 0.033495

is 2.5(KQ/σ0 )2 < t, ai, (w-ai) and h No No No No

Pmax / PQ 1.045 1 1.024375 1
Pmax / PQ < 1.1 Yes Yes Yes Yes
Valid KIC? No No No No

Table 3: Experimental and Theoretical Fracture Toughness

Experimental 𝐾𝑄 Theoretical 𝐾𝑄 Percent
(MPa √𝑚) (MPa √𝑚) Difference (%)
1.0 inch Non-
Fatigued 58.141 29 100.4862
1.5 inch Non-
Fatigued 75.895 29 161.7069
1.5 inch
Fatigued 37.37 29 28.86207
2.0 inch
Fatigued 32.633 29 12.52759
DISCUSSION

As mentioned earlier in the report, the goal of this experiment is to analyze how aluminum
behaves under tensile loading depending on its thickness and whether it is fatigued or not.

It is assumed that a thinner specimen would handle a low amount of stress, and a specimen
subjected to more fatigue would fail quicker than a normal sample. This would also mean that a
specimen with less fatigue and imperfections is expected to handle a higher stress concentration.
After initial measurements, it is observed that ai for a fatigued specimen is greater than that for a
non-fatigued specimen (see Table 1), at 26.17 mm > 22.86 mm.

Based on the results, we can observe for the 1-inch non-fatigued aluminum specimen a normal
pattern for the load vs. displacement curve. It starts with a linear elastic region, and as the
displacement increases, it starts to plastically deform up until fracture at approximately 42000 N
and 2.4 mm, followed by a rapid decrease in load until the specimen is torn apart. Like the 1-
inch specimen, the 1.5-inch non-fatigued specimen showed a similar curve pattern; however, the
fracture was done at a much higher value of loading and displacement at approximately 62000 N
and 2.7 mm. This indicates that our assumption is true, and as the thickness of a material
increases, the material becomes more fracture-resistant and can withstand higher loads for a
certain amount of time.

Moving on with the fatigued specimens, the load vs. displacement curve for the 1.5-inch fatigued
specimen also showed a similar pattern to the previous curves, but the fracture was done sooner
at approximately 41000 N and 2 mm. Similarly, the 2-inch fatigued specimen had the same
pattern as all the curves mentioned above, but the fracture was done even with even lower loads
than the 1.5-inch fatigued specimen at approximately 32000 N and 1.8 mm. After comparing the
fatigued and non-fatigued load vs. displacement graphs, we can conclude that our assumption
were true and that fatigue has a very big impact on the fracture toughness of the material; it
decreases the ability of the material to withstand higher values of loads and makes it even more
brittle.

It is difficult to compare our fracture toughness values with published values for 6061-T6
Aluminum because it was determined that the obtained experimental values for KIC were invalid
as they did not adhere to both criteria. The invalid KIC or KQ values were determined to be
58.141 MPa*m1/2, 75.9 MPa*m1/2, 37.37 MPa*m1/2, and 32.633 MPa*m1/2. Comparing these
values to the published value of 29 MPa*m1/2 [3] it becomes apparent that the experimental
fracture toughness values are widely different with percent errors going all the way up to 161.7%
from the 1.5-inch non-fatigued block found in Table 3. The only experimental fracture toughness
that nears an acceptable difference is that of the fatigued 2.0-inch block which was determined to
have a percent error of 12.52%, which is also found in Table 3.

When looking at the form of fracture of non-fatigued specimens it's usually very smooth and
there’s a lot of elongation which is because the failure is ductile, where we can also observe
either dimples or cup and cone patterns at the fracture site. In addition, the fracture surface has a
uniform appearance and there are no signs of crack propagation. Looking at Figures 5, 6, 9, and
10, it's clear that the crack for the fatigued specimens is uniform and they both showed dimples
at fracture which is, as mentioned earlier, a sign of a ductile fracture. This means that non-
fatigued material fractures in a ductile manner which corresponds to type 1 tearing of the P-v
curve because this curve represents more of a ductile behavior which is following both load vs
curve diagrams and the appearance of the fracture in the figures listed above.

Regarding the visual fracture characteristics of a fatigued specimen surface features like cracks
that initiate at stress concentration and move perpendicular to the direction of applied stress, and
the presence of marks on the surface of our fatigue fracture, called Beach marks, and our crack
propagates from the tip that is an extremely sharp notch or stress concentration. In Figures 13
and 17, we observe both specimens having beach marks, with horizontal lines visible on the
surface of the failed specimen, as well as a sharp initial propagation of the crack, indicating a
brittle and catastrophic failure. This is also backed up by our previously mentioned failure load
value of 40000N for the fatigued specimen, and 62000 N for the non-fatigued specimen, proving
that a fatigued specimen fails at a lower load, with a type 3 failure graph, where the P-V graph of
the 1.5 and 2-inch fatigued specimen have graphs that align with the expected type 3 cleavage
(plane strain failure) and PQ = PMax.
SOURCES OF ERROR

The experiment uses the crack opening displacement gauge (C.O.D gauge) to measure the
displacement of the crack of the specimen. If the C.O.D. is not loaded properly into the crack,
then the displacement values measured during the loading will differ. If the C.O.D. is loaded
deeply into the crack, and then during loading begins to slip to the opening of the crack, the
measured displacement will be larger than the actual displacement of the crack. In the
experiment, it is essential to identify the type of load-displacement behavior to calculate the
stress intensity factor of the specimen. Having inaccurate displacement readings would affect the
identification of the correct load-displacement type of the specimen and would lead to incorrect
calculations and the ability to verify that the specimen has a valid 𝐾𝐼𝐶 .

It also needs to be noted that not all specimens may be fatigued the same. One 1.5-inch fatigued
specimen crack could produce a type 1 load-displacement type, while another could produce a
type 3 due to the difference in the amount of fatigue added. A more fatigued specimen would fail
with a plane strain fracture (type 3) since it is weaker and would fracture at a faster rate. A less
fatigued sample would possibly have a plane stress fracture (type 1) due to the larger difficulty of
fracturing the crack. The variance in fatigued samples could influence the final results for the
fatigued specimen and give it a lower or higher 𝐾𝐼𝐶 . Although the non-fatigued specimens would
not have this source of error, it is possible that imperfections in the aluminum samples also
influenced the results of the non-fatigued specimens. This however is not exclusive to the non-
fatigued specimens as the fatigued specimens could also have results influenced by their
imperfections.

Another possible source of error can be found in the DC-LVDT. This is because the output of
information uses an electrical signal that receives information from the crosshead. The crosshead
is attached to the upper grip which pulls against the aluminum specimen. The crosshead is also
connected to a dual-column frame with a ball screw in each column. Due to the complexity of
this setup, it is possible the connections throughout the setup are not entirely calibrated properly
which can lead to slightly skewed results. Because each of the mechanisms are interconnected
the error should occur throughout the reading which could explain why our load versus
displacement curves produce the expected results when comparing aluminum specimens, but
also produced results that lead to invalid fracture toughness values.

It was determined that the values obtained for fracture toughness were invalid. Comparing these
values with the published fracture toughness for 6061-T6 aluminum determined a percent error
of 105% for 1.0-inch non-fatigued aluminum, 162% for 1.5-inch non-fatigued aluminum, 32.8%
for 1.5-inch fatigue aluminum, and 10% for 2.0-inch fatigued aluminum.
CONCLUSION

Through experimentation, it was determined that none of the four specimens passed the criteria
to have a valid plane strain fracture toughness, 𝐾𝐼𝐶 . The specimens used were relatively thin, and
it was determined that a larger thickness would be necessary to obtain valid values for 𝐾𝐼𝐶 .
Further analysis determined that with increasing material thickness, fracture-resistance also
increases, being able to withstand higher loads before fracture occurs. Fatigue was determined to
have a significant impact on the fracture toughness of the material by decreasing the ability of
the material to withstand higher values of loads and making it even more brittle. Non-fatigued
aluminum specimens were found to fracture in a ductile manner corresponding to type 1 tearing
of the P-v curve. The fatigued aluminum however was found to align with the type 3 cleavage of
plain strain failure due to brittle and catastrophic failure.
The non-fatigued specimens showed a larger variability compared to the fatigued specimens and
produced larger deviations compared to the expected values.

REFERENCES

[1] Tensile testing machines | an introduction. (n.d.-b).

https://www.instron.com/en/resources/test-types/tensile-test

[2] Concordia University, MECH 321 Laboratory Manual 2024, 2024. Pages 50-57.

[3] ASM material data sheet. (n.d.).

https://asm.matweb.com/search/SpecificMaterial.asp?bassnum=ma6061t6
APPENDICES

Using a 1.0-inch Non-Fatigued sample as an example.

𝑚𝑡 obtained using the elastic region of data and then taking the linear slope; as seen in Figure 3.
𝑚5 = 𝑚𝑡 ∗ 0.95
𝑚5 = 17086 ∗ 0.95 = 16231.7
𝑃5 , 𝑃𝑚𝑎𝑥 , 𝑎𝑛𝑑 𝑃𝑞 found using the load-displacement behaviour chart; Figure 2. The points were
taken using Figure 4.

𝑎𝑖
𝛼= 𝑤
22.86𝑚𝑚
𝛼 = 50.225𝑚𝑚

𝛼 = 0.455

𝑃𝑄 2+𝛼
𝐾𝑄 = (𝑡 )(0.866 + 4.64𝛼 − 13.32𝛼 2 + 14.72𝛼 3 − 5.6𝛼 4 )((1−𝛼)1.5 )
√𝑤

40578.96𝑁
𝐾𝑄 = (0.0254𝑚√0.050225𝑚)(0.866 + 4.64(0.455) − 13.32(0.455)2 + 14.72(0.455)3 −
2+0.455
5.6(0.455)4 )((1−0.455)1.5 )

𝐾𝑄 = 59.45 𝑀𝑃𝑎 √𝑚

2
𝐾
2.5 ( 𝜎𝑄 )
𝑦

2
59.45 𝑀𝑃𝑎 √𝑚
2.5 ( )
276 𝑀𝑃𝑎

0.116 𝑚

𝑃𝑚𝑎𝑥
𝑃𝑄

41464.96𝑁
40578.92𝑁
=1.021