Laser Cutting Effects on Fatigue of Austenitic Stainless Steel

Engineering Fracture Mechanics xxx (2017) xxx–xxx
Contents lists available at ScienceDirect
Engineering Fracture Mechanics

journal homepage: www.elsevier.com/locate/engfracmech
Influence of surface condition due to laser beam cutting on the

fatigue behavior of metastable austenitic stainless steel AISI 304
D.F. Pessoa a,b,⇑, P. Herwig b, A. Wetzig b, M. Zimmermann a,b
a
Institut für Werkstoffwissenschaft, Technische Universität Dresden, 01069 Dresden, Germany
b
Fraunhofer-Institut für Werkstoff -und Strahltechnik IWS, 01277 Dresden, Germany
a r t i c l e i n f o a b s t r a c t
Article history: Austenitic steels are among the stainless steels the most common used type of material to
Received 17 November 2016 produce sheet metal parts which can be manufactured by laser beam cutting in a useful
Received in revised form 25 May 2017 manner. However, the use of this technique to manufacture structural thick parts is
Accepted 26 May 2017
restricted due to the creation of macroscopic defects and the lack of reliable fatigue
Available online xxxx
strength data. In this context, the fatigue behavior of sheet-like samples made of meta-
stable austenitic stainless steel AISI 304 cut by laser beam is discussed in this paper. The
Keywords:
analyses showed that laser beam cutting creates three kinds of macroscopic defects, which
Fatigue behavior
Laser cutting
are a pronounced relief-like structure along the cut surface, burr in the underside of the cut
Metastable austenitic stainless steels edge and pores in the interface between the recast layer and base material or inside the
recast layer. As a consequence, the fatigue strength of parts cut by laser beam is 40% lower
in comparison to specimens in a macroscopically quasi defect-free state. An evaluation of
the fatigue results based on comprehensive fractographic analyses allows to explain the
reasons for distinct differences between the aforementioned influence factors. The most
significant reduction of fatigue life is attributed to the notch effect of the burr, followed
by the notch effect created by pores, while the influence of the surface relief is of minor
significance.
Ó 2017 Elsevier Ltd. All rights reserved.
1. Introduction
Laser beam cutting is a thermal cutting process frequently used to manufacture metallic materials. This process is an ide-
ally suited cutting method whenever a low and very localized heat input and the production of components with surfaces
with better quality and free of mechanical deformation is required. A variety of geometries such as sheets of different thick-
nesses or 3D-shaped parts can be cut by this method. Applications range from 0.3 mm thick metallic multi-layer sealing gas-
kets for cylinder head engines to 32 mm thick plates for farm and earth-moving equipment [1]. However, the use of this
technique to manufacture structural thick parts is limited due to the creation of a pronounced surface relief. Furthermore,
to control the influence of the laser cutting on the cyclic strength the interaction between process parameters and geomet-
rical and microstructural modifications in the region of the cut kerf must be identified. During laser cutting different mech-
anisms become operative. Basically, the workpiece is locally heated by radiation and part of the metal is melted. If inert gas
assists the process a narrow penetrating cavity is formed, and the hot molten material is flushed away by the shearing action
⇑ Corresponding author at: Institut für Werkstoffwissenschaft, Technische Universität Dresden, 01069 Dresden, Germany.
E-mail addresses: davi.pessoa@mailbox.tu-dresden.de (D.F. Pessoa), patrick.herwig@iws.fraunhofer.de (P. Herwig), andreas.wetzig@iws.fraunhofer.de
(A. Wetzig), martina.zimmermann@tu-dresden.de (M. Zimmermann).
http://dx.doi.org/10.1016/j.engfracmech.2017.05.040
0013-7944/Ó 2017 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Pessoa DF et al. Influence of surface condition due to laser beam cutting on the fatigue behavior of meta-
stable austenitic stainless steel AISI 304. Engng Fract Mech (2017), http://dx.doi.org/10.1016/j.engfracmech.2017.05.040
2 D.F. Pessoa et al. / Engineering Fracture Mechanics xxx (2017) xxx–xxx
of a coaxial jet. This mechanism, the so-called inert gas melt shearing, on the one hand produces a high surface quality, on
the other hand the cutting speed is not as fast as using an active gas [1,2]. In order to rank the efficiency of the cutting pro-
cesses, the parameter cutting performance (S) is often used [3]. It takes into consideration the cutting speed (v), plate thick-
ness (t), kerf width (k) and surface roughness (R), as follows:
vt
S¼ ð1Þ
kR
Laser beam cutting stands out from other thermal cutting methods by its lower and very localized energy input, causing
smaller layers of recast material and lower width of heat affected zone [4]. In comparison to conventional mechanical cutting
processes, the advantages are high dimensional accuracy, reduced times for setup, no mechanical deformation at cut edge
region, high material utilization rate and high flexibility. However, by increasing the material thickness an increasing surface
roughness along the cut edge can be observed [2], which, in turn, can negatively affect the mechanical properties, in partic-
ular the fatigue strength [5,6]. Nonetheless, the fatigue test results reported by Thomas [7] show that high strength steels cut
by laser beam can have similar fatigue strength in comparison to the smooth parts based on the optimization of process
parameters. Furthermore, the parts cut by laser show better fatigue behavior than those cut by oxy-fuel, plasma and stamp-
ing [4,8]. The complexity of the process becomes apparent when looking at many publications on laser cutting in the context
of the material chemical compositions [7,9,10], cut process parameters such as cutting speed [1,7,11,12], laser output power
[12,13], assist gas pressure [1,12], stand-off between workpiece and nozzle [2], nozzle geometry [1,2,14] and focus position
[15,16]. It is shown, for instance, that a high cutting speed and a low output power produce less pronounced surface relief.
Nonetheless, using this process strategy cut edges with inhomogeneous characteristics are produced [17]. For the metastable
austenitic steel AISI 304 the use of fiber laser leads to the creation of a high surface roughness for sheets thicker than 4 mm.
However, using CO2 laser cutting this behavior is only observed for plates thicker than 8 mm [18]. In addition to the laser
cutting process, the material composition and the prior surface state play an important role on the cutting quality [9,10].
For instance, for steel plates with thickness of 25 mm cut by flame cutting, it is shown that Cu and Ni in low content levels
have a favorable effect on the cut surface characteristics.
Looking at the topic from a manufacturing point of view, the main objective regarding the laser cutting process is to pro-
cess a material as fast as possible producing a surface with good quality concerning roughness and squareness according to
standards mandatory for the targeted application [19]. Nevertheless, to minimize the influence of laser beam cutting on the
cyclic strength, the interaction between the process parameters and the geometrical and microstructural changes in the
region of the cut kerf must be identified. Due to a lack of reliable fatigue strength data for parts cut by laser, the use of this
process to manufacture structural thick parts is restricted, limiting the potential application of this manufacturing method in
such cases.
The fatigue life of samples and components depends not only on the applied stress level, but also substantially on the
initial state of the material (heat-treatment, pre-deformation, defect-afflicted or quasi defect-free), the configuration of
the defect (macroscopic notch and/or microstructural inhomogeneity) and, last but not least, the surface quality. The assess-
ment of these influencing factors in the high to very high cycle fatigue regime must be separately evaluated in order to iden-
tify the dominant failure-relevant effect and, thereby, to accomplish the basis for the material or strength-oriented process
optimization [20].
Laser beam cutting produces geometrical inhomogeneities on the sheet surface, which, in turn, act as a notch effect,
which has a negative influence on the fatigue strength. This localized stress concentration changes the stress distribution
along the part cross section and can act as a crack initiation site. In order to quantify the severity of the notch effect, the
stress concentration factor (Kt) can be used. Kt is defined as the quotient between the maximum stress in the notch root
and the normal stress. When the method applied to analytically calculate the increase of stress is solely based on the
assumption of a purely linear-elastic deformation behavior, the stress concentration factor does not depend on the material
properties, being only a function of the part geometry and loading mode [21]. However, assuming elastic-plastic deformation
near the notch root, the stress concentration factor does no longer apply since the effective notch effect is influenced by local
plastification. For parts cyclically loaded, the discrepancy between Kt and the real stress at the notch root is considered by
using the concept of fatigue effective stress concentration factor (Kf) [21]. This approach is based on experimentally obtained
fatigue results for unnotched and notched specimens. It is defined according (2), where rf denotes the fatigue limit of
unnotched specimens and rnf the fatigue limit of notched specimens for a distinct number of lading cycles.
rf
Kf ¼ ð2Þ
rnf
Due to its good deformability the austenitic steels are among the stainless steels the most common used type of material.
Moreover, under certain conditions this material undergoes a phase transformation from the c-austenite to the harder a0 -
martensite [22], which can have a beneficial effect on the fatigue behavior [23] and can be used to locally optimize the static
and cyclic strength properties of sheet metals by means of the production process [24]. Additionally, to promote the a0 -
martensite formation a critical threshold value of plastic strain amplitude and a certain amount of accumulated plastic strain
needs to be exceeded [25].
D.F. Pessoa et al. / Engineering Fracture Mechanics xxx (2017) xxx–xxx 3
In this context, the purpose of the present study is to investigate the impact of geometrical surface characteristics on the
fatigue behavior for thick parts made of metastable austenitic stainless steels cut by laser beam, thereby, supporting the cut-
ting process optimization concerning fatigue strength.
2. Material, specimens conditions and experiments
In order to evaluate the influence of the laser cutting process on the fatigue behavior, flat specimens made of metastable
austenitic stainless steel type AISI 304 were cut out by high-power solid-state disk laser from sheets with 4 mm and 6 mm
thickness. The geometries of the specimens are depicted in Fig. 1a and b. The plates used to produce the samples were rolled
and supplied in solution-annealed state. In order to exclude any influence from the chemical composition of the material all
samples with the same thickness were produced from the same sheet and therefore from the same batch. The cutting
machine used is a Trumpf TruDisk 5001 laser source, and the most relevant process parameters used to cut the fatigue spec-
imens are listed in Table 1. The definition of process parameters was based on previous efforts for process optimization to cut
the investigated material with the best cut kerf quality regarding surface roughness. These parameters are divided into three
groups: cutting process, assist-gas and laser beam. The ‘‘cutting process” group contains the parameters cutting speed and
focal position, which play a significant role regarding the cutting quality [2,7,11,12]. For the present study, it was determined
that for the focal position ‘‘0” the laser beam is focused on the upper surface of the workpiece. Based on the laser output
power and beam diameter at focal position, the power density during the cutting for both plate thicknesses is
8.04 106 W/cm2. For this reason, to cut the thicker specimens it was necessary to use a cutting speed more than two times
slower. According to the parameters listed in the second group of process parameters, N2 was used as assist-gas. This inert
gas was chosen instead of O2 in order to produce a cut surface with better quality and to avoid the formation of pronounced
dross at the lower side of the cut kerf. For O2 assist-gas, due to the limited exothermic oxidation of iron and chromium in the
molten metal caused by the Cr2O3 film formation on the material surface when it is exposed to oxygen, the material is
shielded from further corrosion. This Cr2O3 film solidifies around the droplets of molten metal during the cutting creating
dross [1]. Moreover, the gas pressure and stand-off nozzle/workpiece play an important role on the cutting quality
[1,2,12]. The parameters from the category ‘‘laser beam” group normally are related to the cutting machine and cannot
be modified. The beam quality M2 characterizes the deviation of the real beam dispersion from an ideal laser beam with
a Gaussian distribution of energy intensity and M2 = 1.
During the cutting process not only a surface relief is created over the cut edge, but also burrs in the form of recast drops
are deposited at the lower side of the cut surface. In order to separately evaluate these effects, fatigue specimens were tested
at three different surface conditions, which are classified in this study as: as-cut condition, without burr condition and pol-
ished condition. The specimens in the as-cut condition did not undergo any post-processing after being cut by laser. For the
specimens without burr the lower sides of the cut edges were mechanically trimmed with two different abrasive papers
coated with SiC in grain sizes of 30 mm and 15 mm. The specimens in the polished condition after being cut by laser were
mechanically ground with three different abrasive papers coated with SiC in grain sizes of 30 mm, 15 mm and 10 mm. Subse-
quently, in order to eliminate any residual stress and martensite coming from the grinding process, the specimens were
electro-chemically polished in a solution with 70 vol.% ethanol, 14.2 vol.% distilled water, 10 vol.% diethylene glycol mono-
Fig. 1. Geometry of fatigue specimens with (a) 4 mm and (b) 6 mm thickness.
Table 1
The most relevant process parameters used to cut the fatigue specimens.
Group Process parameter Plate thickness

4 mm 6 mm
Cutting process Cutting speed [m/min] 5 2.25
Focal length [mm] 200 200
Focal position [mm] 2.5 2.5
Beam diameter at focal position [mm] 218 218
Assist-gas Type of gas Nitrogen Nitrogen
Gas pressure [bar] 14 14
Gas flow [m3/h] 63 63
Nozzle shape Conical convergent
Nozzle internal diameter [mm] 4 4
Stand-off nozzle/workpiece [mm] 0.5 0.5
Laser beam Wave length [mm] 1.04 1.04
Beam quality, M2 10 10
Output power [kW] 3 3
butyl ether and 5.6 vol.% perchloric acid. The electro-chemically polished condition represents the fatigue behavior of the
base material in a macroscopically quasi defect-free state. It is important to point out that this polished condition is generally
not the condition of real parts, which often contain defects on the surface caused by the manufacturing process, e.g. small
pitches from the rolling process. The surface characteristics of the cut edge for samples with 4 mm and 6 mm thickness were
qualitatively evaluated by means of an optical microscope VHX-5000 model from the company Keyence. In order to quantify
the surface relief features, roughness measurements were performed over a length of 12 mm by means of a contact pro-
filometer type Perthometer S3P. Additionally, the chemical compositions were assessed by spectrometry using an analytical
analyzer model Spektromaxx from the company Ametek Materials Analysis Division.
Fatigue tests were performed by means of a resonant pulsation test systemTestronic model from the company Russen-
berger Prüfmaschinen AG, which for the given sample geometries allows for cyclic testing at frequencies around 100 Hz.
All cyclic experiments ran at tensile-tensile loading condition with stress ratio R = 0.1. Moreover, the specimens were cooled
during the fatigue tests by compressed air. Temperature measurements by means of a pyrometer model Marathon MM-LT
from the company Raytec at the beginning of the test series showed that no undesired temperature increase occurred due to
the high frequency. A change in test frequency of Df > 2 Hz was applied as criterion for test interruption. Specimens which
did not exhibit external cracks until the ultimate number of cycles N = 107 were evaluated as run-outs. The fatigue limits
were determined by staircase method according Dixon and Mood [26] analysis using an incremental stress of maximal stress
(rmax) = 20 MPa. The fractographic and microstructure investigations of the fatigued samples were performed with a scan-
ning electron microscope (SEM) type JSM 6610 from the company JEOL and by means of an optical microscope MEF4M
model from the company Leica. In order to assess the initial mechanical properties, tensile tests and hardness measurements
were performed. The machine used to monotonically load the samples is a servo-hydraulic test system model 8501 from the
company Instron. The tests were performed at room temperature at a strain rate of 0.166%/s. The hardness was assessed by
means of a Vickers hardness tester model AMH-43 from the company Leco.
3. Results and discussions
3.1. Specimens and material characteristics
The surface relief, the schematic representation of the roughness measurements positions, the cut edge surface profiles
measured at positions 1, 2 and 3, as well as the burrs of samples with 4 mm and 6 mm thickness are shown in Figs. 2 and 3,
respectively.
One can observe three regions with distinct morphologies along the cut surfaces. At the upper area it is possible to see
finer striations vertically oriented. At the lower region wider striations with more irregular pattern are observed. Addition-
ally, between these two regions a transition area can be recognized. Moreover, the differences between the burr character-
istics for plates with 4 mm and 6 mm thickness are clearly visible. For plates with thickness of 4 mm, one can observe a
continuous dross along the lower side of the cut edge, whereas for 6 mm thickness it appears in the form of separated drops
with 0.30 ± 0.06 mm height periodically spaced in a distance of 0.9 ± 0.5 mm.
Table 2 presents the surface roughness measurements of the cut surface in terms of the average roughness Ra, average of
the difference between the five highest peaks and five deepest valleys Rz and the maximum peak-valley distance Rmax over
the measured length. The measurements were performed at three different positions: at 0.3 mm from the lower side and
upper side (position 1 and 3, respectively) and at the center line of cut edge (position 2). The results show that increasing
the plate thickness results in a more pronounced roughness, and it is not constant over the kerf width, becoming rougher
from the upper side towards the lower side of the cut surface (from position 1 towards position 3).
Fig. 2. (a) Surface relief and the schematic representation of the roughness measurements positions (see lines denominated with 1, 2 and 3), (b) cut edge
surface profiles and (c) burrs of a 4 mm sheet thickness cut by laser.
Fig. 3. (a) Surface relief and the schematic representation of the roughness measurements positions (see lines denominated with 1, 2 and 3), (b) cut edge
surface profiles and (c) burrs of a 6 mm sheet thickness cut by laser.
Table 2
Roughness measurements of the cut edge at three different positions (see Figs. 2 and 3 for the positions 1, 2 and 3) for specimens with 4 mm and 6 mm
thickness.
Specimen Position Ra [mm] Rz [mm] Rmax [mm]

4 mm thick 1 4.1 20.8 22.9
2 5.1 29.7 35.8
3 5.3 32.6 42.4
6 mm thick 1 3.0 18.2 21.6
2 5.0 27.5 33.3
3 6.1 35.1 55.7
The process of cutting and flush melted material away involves complex dynamic mechanisms, which are not laminar and
uniform. According to Figs. 2 and 3, the cut surface presents striations with a repetitive pattern, suggesting that this process
has a cyclic behavior and the material is not continuously melted. Based on observations of the cut edge during the cutting
process using a high speed imaging camera, the laser beam melts a portion of the material, which is expelled by the assist-
gas. Afterwards, a new portion of the material located further forward is melted and expelled, and this process continuously
takes place forming overlaps of the recast material. Moreover, the pattern of these striations becomes more inhomogeneous
towards the lower part of the cut edge. This behavior is promoted by an avalanche-like mechanism based on the fact that the
lower regions of the cut surface accumulate the melted material from the upper regions. Therefore, with more amount of
melted material in the lower area, the roughness becomes more pronounced. Another factor which contributes to the higher
roughness is the fact that the assist-gas loses pressure along the thickness of the cut piece, reducing its efficiency to expel the
melted material from the cut kerf. The same mechanism explains the higher roughness for thicker parts. More energy is nec-
essary to cut the plate, which is promoted by slower cutting speeds. It increases the amount of melted material in the cut
front, which, in turn, influences the increase of roughness.
The maximum produced cutting kerf width using the process parameters presented in Table 1 for the plate with 4 mm
thickness is 0.25 mm. Considering Ra measured at position 3 as the surface roughness R in (1), it results in a cutting perfor-
mance of 2.5 1051/s. For the plate with 6 mm thickness the values of cutting kerf width and cutting performance amounts
to 0.35 mm and 1.1 1051/s, respectively. Thereby, one can postulate a cutting performance more than two times lower to
cut the plates with 6 mm thickness compared to the process performance for 4 mm plates.
The most relevant constituents of the chemical composition of the investigated material are presented in Table 3. The
elements of Fe-Cr-Ni alloys influence the tendency for spontaneous formation of austenite. For the investigated material
the high content of nickel promotes the formation of austenite. Furthermore, the presence of carbon and manganese also
contributes to the austenitic microstructure [27].
Fig. 4 schematically depicts a sketch of a specimen cut by laser beam and its main characteristics in the as-cut condition.
Additionally, the figure shows the stress axis for fatigue tests and the plane named in the present study as ‘‘plane N”.
The scanning electron microscopy analysis of the etched material from a section normal to the rolling direction in the
near-surface area of the laser cut kerf is shown in Fig 5. one can observe austenitic structure with equiaxed grains. The aver-
age grain size determined by linear intercept method without considering the twin grain boundaries is around 26 mm for the
4 mm and 6 mm thick specimens. One can observe a slightly higher density of inclusions in the specimen with 4 mm thick-
ness. Moreover, the recast layer seems to be more pronounced for the 4 mm sheets.
The temperatures Ms (temperature below which the change of phase from austenite to a0 -martensite spontaneously
occurs), Md30 (temperature at which 50 vol.% is transformed to a0 -martensite under a total strain of 30%), and the stacking
fault energy cSF depends on the chemical composition of the material. They are relevant parameters to characterize the sta-
bility of the austenitic phase. The respective calculated values according to the prediction models proposed by [22,28,29] are
presented in Table 3. Hence, a relatively stable austenitic phase at room temperature for low levels of plastic deformation can
be assumed.
Table 3
The most relevant constituents of the chemical composition in weight-%, Ms, Md30 and cSF for specimens with 4 mm and 6 mm thickness.
Thickness Fe Cr Ni Mn Si Cu N C Ms [°C] Md30 [°C] cSF [mJ/m2]

4 mm 69.20 19.56 8.43 1.70 0.52 0.11 0.05 0.03 226.8 9.2 24.76
6 mm 69.70 19.49 8.35 1.38 0.42 0.10 0.07 0.03 238.9 4.4 25.30
stress axis
Fig. 4. Sketch of a specimen in the as-cut condition schematically showing its main characteristics, the plane N and the stress axis for fatigue tests.
a) b)
cut kerf
cut kerf
recast layer recast layer
Fig. 5. Microstructure in the near-surface area of the cut edge oriented according the plane N (see Fig. 4) from a sample with (a) 4 mm and (b) 6 mm
thickness normal to the rolling direction.
Table 4
Mechanical properties for specimens with 4 mm and 6 mm thickness.
Thickness Rp0.2 [MPa] Rm [MPa] eRup % Hardness [HV 0.1]

4 mm 322.6 ± 7.6 642.1 ± 8.2 55.6 ± 0.4 179.3 ± 14.1
6 mm 290.2 ± 5.3 621.1 ± 20.5 56.1 ± 0.5 164.6 ± 15.8
The mechanical properties of the evaluated material for the two different specimen thicknesses are summarized in
Table 4. One can observe the higher yield stress, maximum stress and hardness for the specimens from plates with 4 mm
thickness. Since the flat plates are produced by rolling process, higher degree of deformation to obtain thinner shapes is
expected. This, in turn, could explain the higher strength and yield stress of 4 mm thick specimens in comparison with
6 mm specimens.
When the workpiece is being cut by laser beam a portion of the material is melted and expelled by the assist-gas from the
cut kerf. A layer of recast material in the near-surface region of the cut edge is created as showed in Fig. 4. This finer recast
structure is observable from the specimen surface to a depth of approximately 21 mm for the plate with 4 mm thickness and
8 mm depth for the plate with 6 mm thickness. It is associated with a hardness increase to 211.5 ± 17.5 HV 0.01 for the 4 mm
thick plate, which is around 18% higher in comparison to the bulk material. Due to the narrow layer of recast material for
samples with 6 mm thickness it was not possible to measure the hardness in this region using the Vickers micro indentation.
The microstructural changes in the region of the cut kerf are determined by thermal input from the laser cutting process. For
the same plate thickness, on the one hand when a higher energy rate is introduced in the workpiece a wider layer of recast
material is created. This phenomenon is promoted by lower cutting speed or higher laser output power. On the other hand,
smaller beam diameter at the focal position produces smaller width of recast material. Moreover, higher temperature
changes gradients, which is promoted by higher flow and higher pressure of the assist gas contributing to the creation of
a finer microstructure of the recrystallized material.
During the cutting process a part of the expelled melted material solidifies at the lower side of the cut edge in the form of
dross. It is easy to identify that the geometric inhomogeneity of a burr, which is not completely connected to the plate,
causes a pronounced macroscopic notch effect introducing a technical crack-like feature; see for this in Fig. 6 the scanning
electron microscopy of the etched investigated material in the near-surface area of the underside of the cut kerf in a section
located in the plane N (see Fig. 4) from samples with 4 mm and 6 mm thickness. Furthermore, when the melted material
a) b)
cut kerf
cut kerf
burr
burr
Fig. 6. SEM micrograph of the etched investigated material in the near-surface area of the underside of a cut kerf in a section oriented according the plane N
(see Fig. 4) from a sample with (a) 4 mm and (b) 6 mm thickness.
cut
kerf
recast
layer
pore
Fig. 7. Optical microscopy of a cross section oriented according the plane N (see Fig. 4) in the near-surface area of the cut kerf and an internal defect in the
recast layer of a sample with 4 mm thickness.
recasts on the cut edge some internal defects can be created in form of trapped gas. It was observed in the analysis by means
of optical microscopy presented in Fig. 7 of a cross section oriented according the plane Nin the near-surface area of the cut
kerf for a sample with 4 mm thickness. The micrograph presented in Fig. 7 depicts a recast layer of a width of approximately
50 mm. Commonly, the observed width of such recast layers for 4 mm thick samples is around 26 mm. The excessive presence
of melted material combined with the turbulence effect of the ejection gas is the likely reason for the creation of pores. More-
over, the presence of this kind of internal defects close to the surface is extremely detrimental to the fatigue strength [30],
locally increasing the stress and acting as potential sites for crack initiation.
3.2. Fatigue behavior and fractographic analysis
In Fig. 8 the cyclic testing results and respective S-N curves of fatigue samples with different thicknesses in the as-cut
condition loaded at R = 0.1 are compared. The horizontal straight lines are specified from the fatigue limits determined by
staircase tests and statistical analyses [26] with 50% of survival probability for a total number of cycles N = 107. Without fur-
ther post-processing, the durability of laser cut components underlies all three effects simultaneously: surface roughness
profile, burr along the lower side of the cut edges and recast layer along the cut surface. In this case, a minor tendency
towards a decrease of fatigue strength with increasing sheet thickness can be observed. The difference in the S-N curve
slopes can be attributed to the limited number of tested samples and can therefore be neglected.
Fatigue tests from different sample conditions allow for an evaluation of the influence of surface roughness and burr at
the lower side of the cut edge in comparison to the original material in a macroscopically quasi defect-free state represented
by the polished samples. The test results from specimens with 4 mm and 6 mm thickness without post-processing, with
trimmed burr and with polished surfaces loaded at R = 0.1 are depicted in Figs. 9 and 10, respectively. The fatigue limit anal-
ysis according to Dixon and Mood [26] for the ultimate number of cycles N = 107 and 50% of survival probability shows for
samples with thickness of 4 mm in the as-cut condition a fatigue limit of 353 MPa. For samples in the polished condition the
fatigue limit is 585 MPa. Thus, for the aforementioned conditions, the laser beam cutting of metastable austenitic stainless
steel AISI 304 leads to a fatigue strength reduction of around 40% for the sheets with thickness of 4 mm. Comparing the test
results of samples with laser cut surface in the as-cut condition with the results from specimens where the burr was
removed, no real improvement of fatigue strength due to the post-processing can be observed. The increase resulting from
the trimming of the burr results in a fatigue limit of 386 MPa, which stands for an increase in cyclic strength of around 9%
compared to the as-cut condition. Despite the fact that for 4 mm and 6 mm thick sheets the trimming of the burr does not
seem to have a significant potential to regain the original fatigue strength of the base material, the assessment of specimens
in different conditions demonstrated that the macroscopic notch effect in form of burr seems to have at least a more detri-
mental effect on early failure of laser cut samples in comparison to the influence caused by the surface relief. The later find-
ing is somewhat surprising considering the surface qualities observed for the different thicknesses (see again Figs. 2 and 3).
For specimens with thickness of 6 mm the fatigue limits are 573 MPa, 384 MPa and 342 MPa for the polished condition,
the condition without burr and the as-cut condition, respectively (see Fig. 10). In comparison to the test results from samples
with a thickness of 4 mm, similar effects for the fatigue strength caused by the laser cutting are observed.
Fig. 8. Fatigue test results and S-N curves of specimens with 4 mm and 6 mm thickness tested in the as-cut condition.
Fig. 9. Fatigue test results and S-N curves of specimens with 4 mm thickness tested in the as-cut, without burr and polished condition.
The fatigue limits regarding the maximal stress (rmax, D) calculated according to staircase tests and Dixon and Mood anal-
ysis [26] for samples with 4 mm and 6 mm thickness tested at R = 0.1 for the ultimate number of cycles N = 107 and 10%, 50%
and 90% of survival probability are listed in Table 5. One can observe a tendency towards a fatigue limit reduction with the
increase of plate thickness. However, when the fatigue limit of polished specimens from the same plate thickness is taken
into consideration to quantify the negative influence of cutting process, one can observe, as showed in Table 6, that the fati-
gue strength reduction for specimens in the as-cut and without burr condition is very similar for both evaluated plate thick-
nesses. Based on that and on the scatter of test results showed in Fig. 8, there is no clear evidence of fatigue life reduction
with the increase of plate thickness.
The assessment of the fractured surfaces of samples in the three evaluated conditions gives valuable information about
the failure mechanisms and the most detrimental effects regarding crack initiation and fatigue strength. Figs. 11 and 12 show
the crack starting regions from fractured surfaces oriented according the plane N (see Fig. 4) analyzed by SEM from speci-
mens with 4 mm and 6 mm thickness, respectively, tested at three different surface conditions. Crack initiation for samples
in the as-cut condition always started at the crack-like notch at the burr, underlining the previously mentioned assumptions
regarding the dominance of the macroscopic notch effect. In contrast, cracks in samples with trimmed burr normally initi-
Fig. 10. Fatigue test results and S-N curves of specimens with 6 mm thickness tested in the as-cut, without burr and polished condition.
Table 5
Fatigue limits regarding the maximal stress (rmax, D) for specimens with 4 mm and 6 mm thickness with different surface conditions tested at R = 0.1 with 10%,
50% and 90% of survival probability.
Thickness 4 mm 4 mm 4 mm 6 mm 6 mm 6 mm
Condition As-cut Without burr Polished As-cut Without burr Polished
rmax, D 10% [MPa] 382 454 615 357 417 611
rmax, D 50% [MPa] 353 386 585 342 384 573
rmax, D 90% [MPa] 324 318 555 328 351 535
Table 6
Fatigue limit reduction compared to specimens in the polished condition with specimens with 4 mm and 6 mm thickness in the as-cut and without burr
condition tested at R = 0.1 considering 10%, 50% and 90% of survival probability.
Thickness 4 mm 4 mm 6 mm 6 mm
Condition As-cut Without burr As-cut Without burr
Reduction of fatigue life considering rmax, D 10% 38% 27% 42% 32%
ated at pores presented in the interface between the recast layer and the base material. In the polished condition failure rel-
evant cracks started primarily from inclusions on or close to the surface. However, for some failed polished samples the crack
origin could not be determined unambiguously. Backscattered electron analysis at the SEM revealed that the inclusions have
a non-metallic composition. However, not all polished samples showed crack initiation from non-metallic inclusions, but
also from slip bands at the surface resulting from the rather high amount of plastic deformation during the fatigue tests.
Notwithstanding an evident difference of surface roughness along the cut edge between specimens with 4 mm and 6 mm
thickness in the as-cut condition, the difference of fatigue strength is not as pronounced as expected, as demonstrated by
Fig. 8. This behavior can be explained by the fact that for specimens in the as-cut condition the surface relief is not the dom-
inating fatigue life factor but the inhomogeneous dross located on the lower side of the cut kerf. From (2), the fatigue notch
factor created by burr (Kf, burr) and the fatigue notch factor created by pores (Kf, pore) can be quantified according to (3) and (4),
respectively. Based on the fatigue limits regarding the maximal stress with 50% of survival probability (rmax, D50%) the Kf, burr
from plates with 6 mm thickness is 1.67, whereas it is 1.66 for specimens with 4 mm thickness. However, the absence of
burrs on the cut edge of specimens thicker than 4 mm does not create a regain of fatigue strength as pronounced as for plates
with 2 mm thickness presented in a previous work [31].The fractographic results indicate that the presence of pores in the
interface between the recast layer and the base material is the restricting parameter. According to (4), the fatigue notch fac-
tor with 50% of survival probability created by pores (Kf, pore) for specimens with 4 mm and 6 mm thickness are 1.52 and 1.49,
respectively. It shows that the detrimental effect created by pores is around 10% less pronounced than that created by burrs.
Therefore, to increase the fatigue life of parts thicker than 4 mm cut by laser beam two remedies are postulated. Firstly, the
a)
cut edge
crack
initiation
region
burr
b)
recast layer
crack
initiation
region
pore
c)
crack inclusion
initiation
region
originally
polished
surface
Fig. 11. SEM micrograph in the crack starting region from fractured surfaces oriented according the plane N (see Fig. 4) of specimens with 4 mm thickness
in the (a) as-cut condition tested at rmax = 370 MPa up to failure at 1.9 106 load cycles, (b) without burr condition tested at rmax = 370 MPa up to failure
at 4.2 106 load cycles and (c) polished condition tested at rmax = 630 MPa up to failure at 1.5 105 load cycles.
laser cutting process should be further optimized in order to minimize the burr at the cut edges. Secondly, pores in the recast
material or in the interface between the recast layer and the bulk material should be avoided, which again will be a goal for
the process optimization.
rmax;D50%polished
K f ;burr ¼ ð3Þ
rmax;D50%ascut
rmax;D50%polished
K f ;pore ¼ ð4Þ
rmax;D50%without burr
According to Wetzig et al. [32], burrs at the lower region of the cut surface can be avoided using beam oscillation at a
frequency of 5 kHz and a displacement angle of 1.6° during the cutting process. An alternative solution presented by Birkett
et al. [33] is to use a gas jet directed from the bottom side of the workpiece. Moreover, for thicker plates made of stainless
a)
cut edge
crack
initiation
region
burr
b)
recast layer
pore
crack
initiation
region
c)
originally
polished inclusion
surface
crack
initiation
region
Fig. 12. SEM micrograph in the crack starting region from fractured surfaces oriented according the plane N (see Fig. 4) of specimens with 6 mm thickness
in the (a) as-cut condition tested at rmax = 340 MPa up to failure at 2.8 106 load cycles, (b) without burr condition tested at rmax = 510 MPa up to failure
at 1.2 105 load cycles and (c) polished condition tested at rmax = 570 MPa up to failure at 1.8 105 load cycles.
steel, positioning the laser beam focus below the surface about half of the material thickness may also contributes to the
reduction of burrs [15]. However, this solution also causes a low power density near the top of the plate, as a consequence
reducing the process speeds that are necessary to realize the throughout cut. Caristan et al. [16] showed that using a bifocal
technology with one focus positioned near the top surface and the other near the bottom of the plate, the reduction of dross
and at the same time higher cutting speeds are feasible. Moreover, applying lower cutting speeds to process thick plates can
cause instabilities and abnormal molten regions, blowouts and pores on the recast material [2]. Based on that, the cutting
parameters, in special from the group ‘‘assist-gas” presented in Table 1, shall be carefully evaluated and optimized to pro-
duce parts free of pores in order to improve the fatigue behavior.
4. Conclusions
The fatigue behavior of sheet-like specimens made of metastable austenitic stainless steel type AISI 304 cut by laser beam
was investigated. It was shown that the laser cutting produces changes on the surface and subsurface region of the cut edge
area. A pronounced relief along the cut surface, dross in the underside of the cut edge and pores in the interface between the
recast layer and base material or inside the recast layer are formed. Specimens cut by laser beam show a fatigue limit around
40% lower in comparison to samples in a macroscopically quasi defect-free state. However, even though rougher surfaces are
created when thicker plates are cut, the decrease of fatigue limit is surprising independent of the specimen thickness. The
most significant reduction of fatigue life is not attributed to the surface relief, but to the notch effect of the burr, followed by
the notch effect created by pores. Moreover, the laser beam cutting changes the material microstructure in the near surface
region of the cut edge. A thin layer of recast material along the cut kerf is created. This layer has a finer structure in com-
parison to the bulk material, which, in turn, is correlated with a hardness increase. This creation of a harder surface can have
a beneficial impact on fatigue behavior counteracting the notch effect from the surface relief introduced along the cut edge
by the laser beam cutting. An additional surface post-processing to trim the burr makes little sense in a manufacturing pro-
cess with high cutting speed and high efficiency as the laser cutting. Hence, only an adjustment of the process parameters
and changes in the cutting procedure can make headway in avoiding or minimizing the formation of dross and pores, offer-
ing the likely perspective of a significant increase in fatigue strength. The present study gives important information to sup-
port the process optimization in order to produce plate-liked shapes parts with higher fatigue strength. This, once achieved,
can make the use of laser beam cut process feasible to produce structural components. The study did not yet focus in detail
on the microstructural changes caused by the laser cutting process (extend of recast zone, phase transformation, etc.). Future
work must clarify in which extend the fatigue strength is influenced by these factors.
Acknowledgements
The present work was developed with the support of the National Council for Scientific and Technological Development
(CNPQ) – Brazil.
References
[1] Caristan CL. Laser cutting guide for manufacturing. Michigan: Society of manufacturing engineers; 2004.
[2] Ion JC. Laser processing of engineering materials. Oxford: Elsevier; 2005.
[3] Gabzdyl J. Quality and performance in laser cutting. In: Proceedings of the 2nd nordic laser materials processing conference, vol. 1. Luleå: Luleå
University of Technology; 1989. p. 161–8.
[4] Cicero S, García T, Álvarez JA, Bannister A, Klimpel A, Meizono A, et al. Fatigue behaviour of structural steels with oxy-fuel, plasma and laser cut straight
edges. Definition of eurocode 3 FAT classes. Eng Struct 2016;111:152–61.
[5] Mateo A, Fargas G, Calvo J, Roa JJ. Influence of laser cutting on the fatigue limit of two high strength steels. Mat Test 2015;57:136–40.
[6] Dittmann F, Häfele P. Influence of the cutting edge condition on the fatigue behavior of steel sheets under strain controlled loading. In: Advances in
materials testing for research and practice, vol. 1. Düsseldorf: Stahleisen; 2015. p. 37–42.
[7] Thomas DJ. Optimizing laser cut-edge durability for steel structures in high stress applications. J Constr Steel Res 2016;121:40–9.
[8] Lara A, Picas I, Casellas D. Effect of the cutting process on the fatigue behaviour of press hardened and high strength dual phase steels. Eng Struct
2013;213:1908–19.
[9] Manohar M. CO2 laser beam cutting of steels: materials issues. J Laser Appl 2006;18:101–12.
[10] Manohar M, Bodnar RL. Effect of steel composition on the laser cutting behavior of 25-mm-thick plates. J Laser Appl 2005;17:211–8.
[11] Orishich AM, Malikov AG, Shulyatyev VB, Golyshev AA. Experimental comparison of laser cutting of steel with fiber and CO2 lasers on the basis of
minimal roughness. Phys Procedia 2014;56:875–84.
[12] Yilbas BS. Laser cutting quality assessment and thermal efficiency analysis. J Mater Process Technol 2004;155–156:2106–15.
[13] Goppold C, Zenger K, Herwig P, Wetzig A, Mahrle A, Beyer E. Experimental analysis for improvements of process efficiency and cut edge quality of
fusion cutting with 1 lm laser radiation. Phys Procedia 2014;56:892–900.
[14] Eltawahni HA, Hagino M, Benyounis KY, Inoue T, Olabi AG. Effectof CO2 laser cutting process parameters on edge quality and operating cost of
AISI316L. Opt Laser Technol 2012;44:1068–82.
[15] Eguchi S, Hayashikawa H, Hongu H. Experimental study on high-speed cutting of thick stainless steel 4 kW CO2 laser. In: International congress on
applications of lasers and electro-optics (ICALEO). Orlando: Laser Institute of America; 2000.
[16] Caristan C, Vaidya V, Peyret FC. Laser beam propagation caustics through bifocal lenses. In: International congress on applications of lasers and electro-
optics (ICALEO). Orlando: Laser Institute of America; 2003.
[17] Thomas DJ. The influence of mechanical and CO2 laser cut-edge characteristics on the fatigue life performance of high strength automotive steels. J
Mater Process Technol 2011;211:263–74.
[18] Stelzer S, Mahrle A, Wetzig A, Beyer E. Experimental investigations on fusion cutting stainless steel with fiber and CO2 laser beams. Phys Procedia
2013;41:399–404.
[19] ISO 9013:2002. Classification of thermal cutting: geometrical product specification and quality tolerances. Geneva: International organization for
standardization; 2002.
[20] Cremer M, Zimmermann M, Christ HJ. High-frequency cyclic testing of welded aluminium alloy joints in the region of very high cycle fatigue (VHCF).
Int J Fatigue 2013;57:120–30.
[21] Pilkey WD, Pilkey DF. Stress concentration factors. 3rd ed. New Jersey: John Wiley and Sons Inc.; 2008.
[22] Marshall P. Austenitic stainless steels: microstructure and mechanical properties. Essex: Elsevier applied science publishers; 1984.
[23] Stolarz J, Baffie N, Magnin T. Fatigue short crack behaviour in metastable austenitic stainless steels with different grain sizes. Mater Sci Eng A
2001;319–321:521–6.
[24] Bollenhagen CM. Deformation-induced martensite transformation by multi-stage deformation and its utilization to optimize the HCF and VHCF
properties of austenitic stainless steel sheets. Siegen: University of Siegen; 2011 [PhD Thesis].
[25] Krupp U, West C, Christ HJ. Deformation-induced martensite formation during cyclic deformation of metastable austenitic steel: influence of
temperature and carbon content. Mater Sci Eng A 2008;481–482:713–7.
[26] Dixon WJ, Mood AM. A method for obtaining and analyzing sensitivity data. J Am Stat Assoc 1948;43:108–26.
[27] EN 10088-1:2014. Stainless steels – Part 1: list of the stainless steels. Berlin: German Institute of Standardization; 2014.
[28] Cook AJ, Brown BR. The constitution of iron-nickel-chromium alloys at 550–800 °C. J Iron Steel Inst 1952;171:345–53.
[29] Pickering FB. Physical metallurgical development of stainless steels. In: Proceedings of the conference on stainless steels. Goetebor: Institute of metals;
1984. p. 2–28.
[30] Murakami Y. Metal fatigue: effects of small defects and nonmetallic inclusions. Oxford: Elsevier; 2002.
[31] Pessoa D, Grigorescu A, Herwig P, Zimmermann M. Influence of notch effects created by laser cutting process on fatigue behavior of metastable
austenitic stainless steel. Procedia Eng 2016;160:175–82.
[32] Wetzig A, Scintilla LD, Goppold C, Baumann R, Herwig P, Mahrle A, et al. New progress in laser cutting. In: Proceedings of the 3rd international
conference on laser and plasma application in materials science, Kolkata, vol. 1. p. 4–6.
[33] Birkett FN, Herbert DP, Powell J. Prevention of dross attachment during laser cutting. In: Proceedings of the 2nd International conference lasers in
manufacturing, vol. 1. Birmingham: IFS Ltd; 1985. p. 63–6.

Laser Cutting Effects on Fatigue of Austenitic Stainless Steel

Diunggah oleh

Informasi Dokumen

Deskripsi Asli:

Judul Asli

Hak Cipta

Format Tersedia

Bagikan dokumen Ini

Bagikan atau Tanam Dokumen

Opsi Berbagi

Apakah menurut Anda dokumen ini bermanfaat?

Apakah konten ini tidak pantas?

Hak Cipta:

Format Tersedia

Laser Cutting Effects on Fatigue of Austenitic Stainless Steel

Diunggah oleh

Hak Cipta:

Format Tersedia

Engineering Fracture Mechanics xxx (2017) xxx–xxx

Contents lists available at ScienceDirect

Engineering Fracture Mechanics

Influence of surface condition due to laser beam cutting on the

2. Material, specimens conditions and experiments

Fig. 1. Geometry of fatigue specimens with (a) 4 mm and (b) 6 mm thickness.

Group Process parameter Plate thickness

3. Results and discussions

3.1. Specimens and material characteristics

Specimen Position Ra [mm] Rz [mm] Rmax [mm]

Thickness Fe Cr Ni Mn Si Cu N C Ms [°C] Md30 [°C] cSF [mJ/m2]

recast layer recast layer

Thickness Rp0.2 [MPa] Rm [MPa] eRup % Hardness [HV 0.1]

3.2. Fatigue behavior and fractographic analysis

Anda mungkin juga menyukai