International Journal of Fatigue 111 (2018) 44–53

Contents lists available at ScienceDirect

International Journal of Fatigue

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

Effect of electrical discharge machining on corrosion and corrosion fatigue T

behavior of aluminum alloys
⁎
Saravanan R. Arunachalama, Sarah E. Galyon Dormana, , Richard T. Buckleya, N. Aidan Conradb,
Scott A. Fawaza
a
SAFE Inc, 3290 Hamal Cir, Monument, CO 80132, United States
b
Department of Engineering Mechanics, United States Air Force Academy, 2354 Fairchild Dr. Suite 2J2, USAF Academy, CO 80840, United States

A R T I C L E I N F O A B S T R A C T

Keywords: Electrical discharge machining (EDM) is often used in the manufacturing of fatigue test coupons because it
Electrical discharge machining allows for the production of complex shapes and sharp starting notches. However, the effect that the machining
EDM process has on the corrosion susceptibility due to the surface alterations is unknown. This work focuses on the
Corrosion fatigue effect the machining method has on the corrosion and corrosion fatigue behavior of aluminum alloys. The study
Environmentally assisted cracking
is aimed at understanding the influence of EDM processes on environmentally assisted cracking evaluations,
namely crack growth rate. Conventional milling and electrical discharge machining were the machining process
examined in this study; the machining parameters used were best practices for laboratory fatigue sample pro-
duction. The following aluminum alloys (AA) 2024-T351, 5083-H116, 6061-T6 and 7075-T651 were evaluated
in this study. EDM was performed using a 0.152 mm (0.006 in.) hard brass wire. Surface quality evaluations,
microstructural analysis, electrochemical tests and corrosion fatigue testing in sodium chloride (NaCl) solution
were completed. The study found corrosion rates are affected by the EDM machining in all aluminum alloys. In
final testing, crack growth rate tests were completed in sodium chloride solution with starter notches that were
traditionally cut and EDM machined for 7xxx and 2xxx series alloys. The testing showed accelerated fatigue
crack growth rates for the samples with EDM notches as compared to cut notches.

1. Introduction state of the surface and subsurface material are greatly influenced by
manufacturing processes. Surface integrity (SI) is defined as the in-
During an outdoor exposure at the Battelle Outdoor Exposure herent or enhanced surface condition of a sample produced during
Facility in Daytona Beach, FL, a 5083-H116 aluminum alloy compact machining or other surface finishing operations [1]. Extensive research
tension sample (C(T)) exhibited corrosion damage near the notch, has been completed on the SI impacts of machining processes on the
Fig. 1. The notch on the sample had been electrical discharge machined mechanical properties and fatigue life prediction of machined products
(EDM), while the other sample surfaces had been milled. Based on a [2,3]. These reviews focus on typical surface alternations that occur
visual examination of the sample, it appeared that the 5083-H116 test during machining such as phase transformations, microhardness and
coupon suffered corrosion damage on the surface adjacent to the EDM residual stress changes from different manufacturing processes. These
notch, while the milled surface was not affected as severely by the surface effects are often correlated to the mechanical performance,
environment. The ellipse in Fig. 1 highlights the accelerated corrosion primarily fatigue life. Examination of the influence of machining
around the EDM notch. Accelerated corrosion around an EDM notch on method on a tool steel, SAEJ438b, revealed a 35% loss in fatigue en-
a laboratory fatigue specimen raises concerns about the impact of durance in the case of an EDM machined specimen compared to con-
manufacturing methods on corrosion performance, and in turn en- ventionally milled specimen [4]. While lower residual stresses were
vironmentally assisted crack growth testing. Such preferential corrosion observed in the EDM machined specimens compared to milled and
damage on a critical location of a test specimen could produce error in ground specimens for steel and γ titanium aluminide (γ - TiAl) alloys;
damage predictions or material properties based on laboratory testing the microstructure alteration and brittle phase formation that occurred
of specimens. during processing appears to have reduced fatigue performance [3]. It
It is well understood that the surface integrity and the metallurgical is clear from these literature case studies that EDM machining can have

⁎
Corresponding author.
E-mail address: [email protected] (S.E. Galyon Dorman).

https://doi.org/10.1016/j.ijfatigue.2018.02.005
Received 5 October 2017; Received in revised form 5 January 2018; Accepted 5 February 2018
Available online 07 February 2018
0142-1123/ © 2018 Elsevier Ltd. All rights reserved.
S.R. Arunachalam et al. International Journal of Fatigue 111 (2018) 44–53

Nomenclature N cycles
NaCl sodium chloride
ΔK stress intensity range OES optical emission spectroscope
a crack length OCP open circuit potential
C(T) compact tension sample R stress ratio
Cu copper Ra surface roughness
EDM electrical discharge machining Rz surface roughness
EDS energy dispersive X-ray spectroscopy SI surface integrity
ESE(T) eccentrically loaded single edge notch tension sample SEM scanning electron microscope
f frequency wt% weight percent
M molarity Zn zinc

a deleterious effect on the fatigue performance when compared to a 6061-T6 and 7075-T651, tested were from a 6.35 mm (0.25 in.) plate
milling process, particularly in cases of stress-life testing. with no clad layer. Rectangular specimens,
Studies on the effect of EDM machining on the mechanical prop- 5.4 mm × 5.4 mm × 102 mm, were used for characterizing roughness,
erties in aluminum alloys are limited. EDM machining evaluations in microstructural changes, hardness and corrosion properties. To com-
aluminum alloys primarily focus on optimizing the EDM process para- pare the manufacturing methods, the parameters used for conventional
meters [4–6]. In the case of aluminum alloy 6061, no microstructural milling and EDM were the best machining practices for the manu-
changes or hardness variation were noted when comparing specimens facturing of corrosion fatigue laboratory test specimens. No effort was
manufactured by EDM to other cutting processes [7]. Another report on made to optimize or control machining effects beyond what would ty-
5083 noted that the microstructure and corrosion characteristics of the pically be done for a fatigue sample. The same machinist manufactured
alloy can be improved by modifying the recast layer chemistry using a all samples to limit personnel variations. EDM was performed using
silicon electrode in the EDM process [8]. The observation of the pre- 0.152 mm (0.006 in.) hard brass wire; the composition was verified to
ferential corrosion attack near the EDM notch, Fig. 1, and the sus- be 40% zinc (Zn) and 60% copper (Cu). The EDM used in this study was
ceptibility of aluminum alloy 5083 to low temperature sensitized stress an Agie Charmilles FI 240CC which uses deionized water as the ma-
corrosion cracking (SCC) due to thermal exposure indicates that there is chining medium. The EDM machine controls the machining parameters
a strong need for understanding the effect of EDM on a range of alu- based on user inputs of the sample geometry, substrate (for example:
minum alloys [9]. This data is crucial to determining if EDM is a sui- aluminum) and wire materials (hard brass). Machining parameters
table machining method for the manufacturing of fatigue specimens for were not otherwise controlled by the machinist.
testing in accelerated corrosion environments. Any alteration in the
microstructure or surface chemistry on the surface or sub-surface of an 2.2. Surface analysis
EDM machined coupon could produce a crack growth rates or corrosion
rates that are not representative of the base alloy. To evaluate the effect The surface roughness for all of the milled and EDM aluminum alloy
of EDM an evaluation of the microstructure, corrosion and corrosion samples was analyzed using a Mitutoyo SJ-400 profilometer. Roughness
fatigue behavior of the following aluminum alloys 2024-T351, 6061- measurements were performed at four different locations on the ma-
T6, 5083-H116 and 7075-T651was completed using samples manu- chined surface and the average values are reported. Two standard
factured by milling and EDM methods. measurements were carried out, an arithmetical mean (average
roughness), Ra, and measurement of the difference between the tallest
“peak” and the deepest “valley”, Rz. The surface topography of the
2. Experimental procedures
machined surface was also analyzed using scanning electron micro-
scopy (SEM).
2.1. Materials and machining conditions

The aluminum alloy compositions evaluated are listed in Table 1. 2.3. Microstructural analysis and hardness measurements
Alloy, temper, hardness and conductivity were verified using a Thermo
Scientific ARL Quanto Desk - Optical Emission Spectroscope (OES), a Metallographic examinations were carried out on a cross section of
Shimadzu micro-hardness tester and a Hockings Autosigma 3000 elec- the milled and EDM specimens to evaluate the microstructural changes
trical conductivity meter. All aluminum alloys, 2024-T351, 5083-H116, using an optical microscope. The specimens were further examined
using SEM to document any alloy chemistry changes that occurred
during the machining process with aid of energy dispersive x-ray ana-
lysis (EDS). EDS analysis was performed on both the machined surface
as well as in the sub-surface of the cross sectioned specimens. The cross
sectioned specimens were also used to measure the hardness through
the thickness of the sample using a MTS SA2 nano-indenter.

2.4. Corrosion experiments

To evaluate the corrosion behavior of the machined specimens, both

quantitative and qualitative tests were performed. Open circuit poten-
tial (OCP) and linear polarization resistance were measured using a
Gamry Reference 3000 potentiostat. A scan rate of 4 mV/s was used for
the polarization resistance testing. A silver/silver chloride (Ag/AgCl)
reference electrode was used for tall testing. The OCP was measured for
one hour, and the last 30 min were averaged to report an OCP value
Fig. 1. C(T) specimen after two months outdoor exposure at Daytona Beach.
[10]. The qualitative analysis of corrosion damage was assessed by fully

45
S.R. Arunachalam et al. International Journal of Fatigue 111 (2018) 44–53

Table 1
Aluminum alloys investigated in this study with hardness and conductivity values. Composition given in weight percent (wt%).

Alloy Alloy chemistry Hardness Conductivity

Si Fe Cu Mn Mg Cr Zn Ti Al VHN % IACS

2024-T351 0.04 0.05 4.16 0.53 1.27 – 0.01 – Balance 133 29.8
5083-H116 0.08 0.3 0.04 0.54 4.76 0.06 0.009 – Balance 98.4 28.8
6061-T6 0.65 0.37 0.28 0.11 1.13 0.21 – – Balance 103 46
7075-T651 0.07 0.18 0.14 0.05 2.16 0.22 5.67 0.02 Balance 175 31.4

immersing each sample for 48 h in 0.06 M sodium chloride (NaCl) so- 21–22% decrease in hardness was observed in the 6061-T6 and 7075-
lution then visually examining the surface. For both of the electro- T651 alloys, whereas 2024-T351 showed a 13% decrease and 5083-
chemical characterizations as well as the full immersion studies, a 0.6 M H116 alloy showed a 5% decrease compared to the bulk hardness va-
NaCl solution with H2O2 added to ensure sufficient cathodic reactant lues. The measurement locations were verified under in the SEM and
was used as the electrolyte [10]. Fig. 10 shows nano-indentions placed within the HAZ.
Elemental distribution within the spatter region, recast layers and
2.5. Corrosion fatigue testing HAZ showed enrichment of copper on all the aluminum alloys. A re-
presentative SEM image with EDS mapping on a 5083-H116 EDM
Corrosion fatigue testing was completed using eccentrically loaded surface is shown in Fig. 11, and the quantitative analysis of elemental
single edge tension (ESE(T)) samples with the notch placed by saw composition is given Table 4. Copper enrichment was more significant
cutting and EDM, Fig. 2. Testing was completed in 0.06 M NaCl solution on the recast features compared to HAZ zone. However, compared to
using a constant ΔK test profile at 5 MPa √m with a stress ratio (R) of the bulk material composition, the copper enrichment in the HAZ is still
0.1, at a frequency of 1 Hz. Constant ΔK testing was used to determine if notably higher. For example, copper is present in the 5083-H116 alloy
there was a length effect related to the notch, in other words if the EDM in trace amounts, 0.04 wt%, whereas, on the recast layer and HAZ, the
notch was affecting the crack growth rate for a certain distance beyond elemental analysis showed as high as 36 wt% and 1.5 wt%, respectively.
the notch, a change in the linearity of the crack length versus cycles (a Likewise the zinc present in the 5083-H116 EDS mapping on the surface
versus N) curve would be noted. The crack growth rate was measured is higher (2.5 wt%) than the base material (0.009 wt%), confirming the
using direct current potential drop (DCPD). Aluminium alloys 7075- composition changes are due to metal transfer from the brass EDM
T651, 7050-T7451 and 2024-T351 were used in the corrosion fatigue electrode.
study.
3.3. Corrosion experiments
3. Results
Corrosion potential (OCP) studies were completed on EDM and
3.1. Surface analysis milled samples. The OCP data for the 5083-H116 alloy is shown in
Fig. 12. OCP measurements were taken continuously for one hour and
The results of the surface roughness measurements are given below the measurements taken in the last 30 min were averaged to determine
in Table 2. Surface measurements were taken from four different lo- the open circuit potential as per ASTM G69 standard [10]. The average
cations “Ra and Rz” are given. Visual as well as SEM observation of the open circuit potential measured for all the aluminum alloys is given in
surface topography confirmed the surface quality of the EDM machined Table 5. All aluminum alloys show a significant increase in the initial
specimens were inferior (rougher) compared to the milled surface for corrosion potential for the EDM samples compared to the milled sam-
all of the aluminum alloys. SEM observation of milled and EDM ma- ples this is expected based on the increase in copper content on the
chined surfaces are shown in Figs. 3 and 4. The milled surface showed surface of the EDM samples [10]. Likewise the OCP value fell over time,
micro-tearing and surface smearing features, whereas, EDM machined likely due to the slowing of corrosion due to oxide formation on the
surface shows craters from electric arc surface melting. A high magni-
fication micrograph of arc craters and the recast surface with small
particles adhering on the re-melt surface are shown in Fig. 5.

3.2. Microstructural analysis and hardness measurements

Metallographic examinations were carried out on the cross section

of the EDM machined specimens to determine the depth of the heat
affected zone (HAZ) for the different aluminum alloy systems. The EDM
process has been documented to create different layers, a spatter, recast Notch Depth= 5mm
141 mm
and a heat effected zone, the total thickness of the layers will be af-
fected by the current used in the EDM process [11]. The metallographic
Sample Thickness= 3.2 mm
analysis showed that the thermally effected layer were approximately
10 µm deep, however it was not always possible to note the boundary
between the spatter and recast layers. The metallographic observation
revealed two layers for all of the aluminum alloys investigated, shown
in Figs. 6–9. Based on the 10 μm process zone, if surface finishing is
completed post EDM the spattered and recast layers would likely be
removed, however the HAZ could remain.
Multiple hardness measurements were taken throughout the HAZ 38.1 mm
and averaged for each sample, Table 3. The HAZ for the EDM specimens
Fig. 2. ESE(T) sample used for corrosion fatigue testing.
showed a reduction in hardness compared to the bulk material. A

46
S.R. Arunachalam et al. International Journal of Fatigue 111 (2018) 44–53

Table 2
Surface roughness values using conventional milling and EDM process. An arithmetical mean (average roughness), Ra, and measurement of the difference between the tallest “peak” and
the deepest “valley”, Rz, are reported.

Machining method 2024-T351 5083-H116 6061-T6 7075-T651

Ra, µm Rz, µm Ra, µm Rz, µm Ra, µm Rz, µm Ra, µm Rz, µm

Milled 1.78 ± 0.2 9.11 ± 0.1 1.51 ± 0.2 7.35 ± 0.1 1.91 ± 0.1 9.02 ± 0.1 2.11 ± 0.1 9.26 ± 0.1
EDM 2.68 ± 0.5 17.2 ± 0.5 2.67 ± 0.4 16.6 ± 0.3 2.54 ± 0.6 16.2 ± 0.3 2.45 ± 0.5 15.7 ± 0.4

Particles

Crater

Fig. 3. SEM image of milled 7075-T651 surface.

Fig. 5. SEM image shows higher magnification EDM 5083-H116 surface with crater, flow
over and re-deposits particles.

Fig. 6. Optical micrograph shows altered metal zone due to EDM in 2024-T351 specimen.

Fig. 4. SEM image of EDM 5083-H116 surface. both milled and EDM surfaces using polarization resistance. Fig. 13
shows a comparison of the milled and EDM data sets for all alloys. The
surface. The final OCP value for the 2024-T351, 5083-H116 and 7075- most prominent distinction in Fig. 13 is the positive shift in the zero
T651 samples was more cathodic than the milled samples. The 6061-T6 current potential and elongation of the ohmic region of the EDM ma-
sample was slightly more anodic, this could be due to the increased zinc chined specimens. Corrosion rates calculated from these studies are
deposition on the surface compared to the base material, which has given in Table 6. The corrosion rate was higher for all aluminum alloys
negligible amounts of zinc (Table 1). All other alloys have zinc present machined using EDM compared to the milled specimens. This data
making that effect less notable [10]. The 5083-H116 and 7075-T651 should be used to compare the corrosion rate between machining
milled samples had OCP values that were more reactive (negative) than methods rather than making comparisons between alloys.
the published values, the reason for this is unknown, but comparisons A visual comparison of the corrosion damage was completed for
can still be made between machining methods for each alloy. each of the alloys with respect to the machining method by a 48 h full
Corrosion rates for all the aluminum alloys were determined from immersion study in 0.6 M NaCl. The visual examinations indicate the

47
S.R. Arunachalam et al. International Journal of Fatigue 111 (2018) 44–53

Fig. 7. Optical micrograph shows altered metal zone due to EDM in 5083-H116.

Fig. 10. SEM micrograph shows indents placed within the HAZ zone for 2024-T351

Fig. 8. Optical micrograph shows altered metal zone to EDM in 6061-T6 alloy.

Fig. 11. SEM image and elemental mapping for copper (Cu) and zinc (Zn) on 5083-H116
EDM surface shows copper enrichment on the recast features

Fig. 9. Optical micrograph shows altered metal zone due to EDM in 7075-T651 alloy EDM surface corroded more severely than the milled surface. As ex-
pected the 5083-H116 alloy showed less corrosion attack compared to
other alloy systems studied. Macrographs of the corroded surfaces of
Table 3
the milled and EDM samples for all alloys are shown in Figs. 14 and 15,
Average hardness values measured in HAZ and bulk material analysis.
respectively. It should be noted that the while the corrosion rates in
Material/layer 2024-T351 5083-H116 6061-T6 7075-T651 Table 6 are lower for the milled 2024-T351 than the milled 5083-H116,
the corrosion damage noted for the 2024-T351 after 48 h is more in line
HAZ hardness (VHN) 116 93.7 81.8 137
with the expected corrosion resistance of the alloys. The reason for the
Bulk hardness (VHN) 133 98.4 103 175
Difference (%) 13% 5% 21% 22%
low corrosion rate for the milled 2024-T351 is in Table 6 unknown,
perhaps if the sample was tested after a longer time a different result
would be noted. The polarization scans are relatively quick tests, the

48
S.R. Arunachalam et al. International Journal of Fatigue 111 (2018) 44–53

Table 4 is also attributed to the commonly observed inferior stress-life fatigue

Quantitative elemental analysis (K-series) from the EDS mapping location on 5083-H116 performance in steels and titanium alloys for EDM machined coupons
shown in Fig. 11.
compared to milled or abrasive water jet machined coupons. The stress-
Element line Element wt% wt% Error life fatigue behavior of any materials is grossly affected by the surface
finish. In addition, the observation of recast layer on the surface and
OK 16.2 ± 0.0 HAZ zone in the sub-surface of EDM specimens is expected to affect the
Mg K 4.0 ± 0.0
fatigue properties [1–3]. However none of those studies take into ac-
Al K 75.4 ± 0.0
Mn K 4.0 ± 0.0 count the effect on crack growth rate which due to the presence of a
Fe K 0.1 ± 0.0 starter notch and precrack would not be affected by surface roughness.
Cu K 1.5 ± 0.0 Based on microstructural characterization the depth of HAZ for all
Zn K 2.5 ± 0.0 aluminum alloys examined was on the order of 10 µm. This size is
Total 100 within the range of published values for other metallic systems that
have been processed by EDM [11,12]. Microstructural changes in the
HAZ were reflected in the reduced hardness values measured in the
samples would not be in solution over 2 h. Aluminum alloy 2024-T351 altered metal layer. The hardness of HAZ in the 5083-H116 was only
is particularly susceptible to intergranular corrosion (IGC) which is a slightly reduce (5% difference) compared to 13–22% reduction in other
time dependent process [16]. alloy system. This small hardness variation in the HAZ can be attributed
to the fact that 5083-H116 is not a heat treatable alloy. Further, ob-
3.4. Corrosion fatigue testing servation of copper enrichment in the recast layer as well as in the HAZ
is an expected phenomenon. EDM electrode material can often influ-
The results of the constant ΔK testing of 7075-T651, 2024-T351 and ence the chemistry of the HAZ. The HAZ layer is affected by micro-
7050-T7451 samples with a notch placed with EDM and by conven- structural damage or alterations due to recast layer, re-dissolution of
tional cutting are shown in Figs. 16–18. All testing was completed using phases and material transfer from the electrode [8,11,12]. Micro-
a constant ΔK of 5 MPa √m, R of 0.1 and a frequency of 1 Hz in 0.06 M structural changes due to re-deposition from the electrode material
NaCl, the final crack length was 10 mm for each sample. The slope of have been reported. Even in the case of 5083-H116, where the material
the crack length (a) versus cycles (N) curve is the crack growth rate for is more prone to SCC sensitization due to heat exposure, published
each test. These results clearly show that the presence of the EDM notch results showed that surface modification of the HAZ using pure silicon
causes an increase in the crack growth rate for all alloys tested; Table 7 electrode in EDM process improved corrosion resistance suggesting the
summarizes the crack growth rates for each alloy. material was not sensitized by the temperatures experienced during the
To further investigate the role of copper in the increased crack EDM process [8]. The, EDM electrode used in this study was hard brass,
growth rate of aluminum alloys with EDM notches, copper tape was 60% copper. During the EDM process copper dissolves in the recast
applied to the notch of 7075-T651 sample and the sample corrosion layer and then defuses in the HAZ affecting the surface corrosion
fatigue test was completed using the constant ΔK protocol. The results properties. A higher corrosion rate was observed in all the aluminum
of the test with the copper tape on the notch surface appear in Fig. 19. alloys machined using EDM compared to the milled surface. It has been
The sample with the copper present in the notch had a crack growth reported that copper ions in concentration of 0.02 to 0.05 part per
rate 24% faster than the cut notch. million (ppm) added to a neutral or acidic solution can cause pitting
corrosion to occur in aluminum alloys [13]. Similar observations have
4. Discussion reported in Al-Mg systems [14]. Thus, corrosion rate changes due to
copper enrichment should be expected when EDM is used with a brass
A primary motivation of this study was to understand the effect of electrode.
EDM on the corrosion fatigue behavior of aluminum alloys. To in- Likewise the corrosion fatigue testing at a constant ΔK of 5 MPa √m,
vestigate this effect various evaluations were completed comparing at an R of 0.1 showed that the presence of an EDM notch increased the
milled and EDM samples. The EDM samples had higher corrosion rates fatigue crack growth rate as compared to a cut notch for 7xxx and 2xxx
compared to milled samples, which was expected based on published alloys. The cause for this increase in crack growth rate is likely the
literature on other metal systems [1–3]. The spattering and crater for- increased copper in the notch acting as a cathodic site near to crack
mation during spark erosion in EDM process is generally considered to mouth. The effect of copper in aluminum alloys well documented from
be the cause of high surface roughness for EDM samples. This roughness studies on pitting and microgalvanic effects of alloying elements in

Fig. 12. Comparison of corrosion potential measurements on a milled and EDM surface of 5083-H116, all measurements were taken with a Ag/AgCl reference electrode.

49
S.R. Arunachalam et al. International Journal of Fatigue 111 (2018) 44–53

Table 5
Summary of average open circuit potentials measured using an Ag/AgCl reference electrode for milled and EDM processed aluminum alloy surfaces. ASM values were originally reported
using a saturated calomel electrode and converted to Ag/AgCl for clarity of comparison 13.

Method/material 2024-T351 5083-H116 6061-T6 7075-T651

ASM13 −652 mV −687 mV −694 mV −685 mV

Milled −648 ± 4 mV −722 ± 3 mV −686 ± 3 mV −750 ± 5 mV
EDM −625 ± 16 mV −693 ± 12 mV −695 ± 11 mV −710 ± 8 mV
EDM (Initial) −548 ± 30 mV −677 ± 31 mV −620 ± 28 mV −655 ± 34 mV

Fig. 13. Comparison of polarization resistance of milled and EDM surfaces of 5083-H116 alloy, all measurements were taken with a Ag/AgCl reference electrode.

Table 6
Corrosion rates (millimeter per year (mm/year)) calculated from polarization resistance.

Material/condition 2024-T351 5083-H116 6061-T6 7075-T651

Milled 0.005 0.014 0.012 0.130

EDM 0.330 0.191 0.203 0.305

Fig. 15. Comparison of corrosion damage for EDM surfaces for all aluminum alloys after
48 h immersion in 0.6 M NaCl solution

is likely the case for the EDM copper enrichment zones, explaining the
increased corrosion damage along the EDM notch from outdoor ex-
posure, Fig. 1.
It was assumed prior to completing the corrosion fatigue testing,
that the fatigue crack would outgrow the effect of the EDM notch
Fig. 14. Comparison of corrosion damage for milled surfaces for all aluminum alloys after
48 h immersion in 0.6 M NaCl solution. during the test. The return to cut notch fatigue crack growth rates was
expected to occur when the crack was approximately the same size or
larger than the spatter and HAZ regions, 10 µm. The results of the
aircraft aluminum alloys, 2xxx and 7xxx series alloys in particular. It is
corrosion fatigue testing at a constant ΔK did not support this ex-
known that copper is needed for the precipitation-hardening effects for
pectation. Rather the crack growth rate remained constant and ac-
these 7xxx and 2xxx series alloys. It is also understood that these copper
celerated for a full 5 mm of crack growth, as shown by the linear re-
containing intermetallic particles (Al2Cu, Al2CuMg, Mg (Zn,Cu, Al)2)),
gression R2 value being above 0.995 for the a versus N plots in
are critical to strengthening the aluminum alloy [16,17]. The presence
Figs. 16–18 [32]. This result shows that the increased corrosion oc-
of copper intermetallics in these alloys systems are also thought to be a
curring at the EDM notch is changing the environment within the crack
driving factor as to why these alloys are susceptible to localized cor-
tip. For crack growth rates in full immersion sodium chloride, the high
rosion such as pitting, crevice corrosion and intergranular corrosion
copper content in the notch is a cathode at the mouth of the crack
(IGC) [15–30]. It has also been hypothesized that copper can aid in self-
allowing for a higher current density between the galvanic couple of the
inhibition of fatigue crack growth rates in Al-Cu-Mg and Al-Cu-Li alloys
aluminum and copper. This galvanic coupling with copper causes the
[31]. It has been shown that copper once it comes out of solid solution
aluminum to corrode at a faster rate in NaCl solution than without the
is detrimental from a corrosion stand point to the aluminum matrix; this

50
S.R. Arunachalam et al. International Journal of Fatigue 111 (2018) 44–53

Fig. 16. Results of the constant ΔK testing in 7075-T651 at 5 MPa √m, R of 0.1 and a frequency of 1 Hz, the blue line is for the sample with the notch placed using EDM, the red line is the
sample with the conventionally cut notch. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

copper present, see Table 6. This galvanic effect can change the fatigue EDM notch and a cut notch sample has broken down due to the in-
crack growth rate in aluminum through either hydrogen effects or creased corrosion processes for the EDM fatigue sample compared to
anodic dissolution [33–36]. For the hydrogen effect, the increase in the the cut sample. The corrosion fatigue testing with the copper tape in a
corrosion rate is indicative of a faster corrosion process, meaning more cut notch confirmed the effect of increased copper in the notch on crack
hydrogen is being produced at the area of high copper. This increase in growth rates. These results with fatigue crack growth rates, also suggest
hydrogen is then in turn entering the crack tip and increasing fatigue that care should be taken when using EDM machining even with an
crack growth rates. For anodic dissolution the galvanic effect increases electrode other than brass for environmentally assisted cracking tests,
the current allowing more aluminum at the crack tip to be dissolved for the electrode may enrich the sample surface with material during the
each cycle than for the cut notch. In either case, similitude between and spark discharge process which means that corrosion rates in the area of

Fig. 17. Results of the constant ΔK testing in 2024-T351 at 5 MPa √m, R of 0.1 and a frequency of 1 Hz, the blue line is for the sample with the notch placed using EDM, the red line is the
sample with the conventionally cut notch. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

51
S.R. Arunachalam et al. International Journal of Fatigue 111 (2018) 44–53

Fig. 18. Results of the constant ΔK testing in 7050-T7451 at 5 MPa √m, R of 0.1 and a frequency of 1 Hz, the blue line is for the sample with the notch placed using EDM, the green line is
the sample with the conventionally cut notch. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Table 7 the notch could be different from the bulk material [8]. If EDM is to be
Results of the constant ΔK testing in three aluminum alloys at 5 MPa √m, R of 0.1 and a used for environmental fatigue sample production an evaluation of the
frequency of 1 Hz, with the notch placed using EDM, or conventionally cut.
process zone should be completed to verify the effect is acceptable. For
Alloy EDM Saw cut Percent difference (%)
environmental fatigue testing the environment is expected to be a
(mm/cycle) (mm/cycle) controlled parameter, therefore any changes to the corrosion reaction
rates outside of the intended material/environment combination can be
AA7075-T651 7.77 × 10−5 4.48 × 10−5 42 detrimental to the results of the experiments.
AA2024-T351 1.19 × 10−5 9.03 × 10−6 24
AA7050-T7451 3.40 × 10−5 2.36 × 10−5 31

Fig. 19. Fatigue crack growth rates (slope of line) for 7075-T651 samples tested in 0.06 M NaCl at a constant ΔK of 5 MPa √m, R = 0.1 and a frequency of 1 Hz. The crack growth rate for
the sample with copper tape applied to the notch is 24% faster than the saw cut notch.

52
S.R. Arunachalam et al. International Journal of Fatigue 111 (2018) 44–53

5. Conclusions [6] Pramanik A, Basak AK, Islam MN, Littlefair G. Electrical discharge machining of
6061 aluminum alloy. Trans Non-Ferrous Met Soc China 2015;25:2866–74.
[7] Akkurt A. The effect of cutting process on the surface microstructure and hardness
Based on the study on the effect of EDM machining using a brass of pure Al and 6061 Al alloy. Eng Sci Technol Int J 2015;15:303–8.
electrode of 2024-T351, 5083-H116, 6061-T6, and 7075-T651 alu- [8] Stambekova K, Lin H, Uan J. Microstructural and corrosion characteristics of al-
minum alloys on the corrosion behavior, following conclusion can be loying modified Layer on 5083 Al alloy by electrical discharge alloying process with
pure silicon electrode. Mater Trans Jpn Inst Met 2012;53(8):1436–42.
drawn: [9] Kramer L, Phillippi M, Tack WT, Wong C. Locally reversing sensitization in 5XXX
aluminum plate. JMEPEG 2012;21(6):1025–9.
• A HAZ layer thickness of 10 µm is found to be common for all the [10] ASTM G69. Standard test method for measurement of corrosion potential of alu-
minum alloys. American Society for Testing and Materials, West Conshohocken;
aluminum alloys.
• Hardness measured on the HAZ is reduced compared to the bulk
2001.
[11] Klocke F, Welling D, Dieckmann J. Comparison of grinding and wire EDM con-
material in 2024-T351, 6061-T6 and 7075-T651 alloys, the hardness cerning fatigue strength and surface integrity of machined Ti6Al4V components.
Procedia Eng 2011;11:184–9.
was only slightly reduced for 5083-H116.
•
[12] Iqbal AA, Khan AA. Influence of process parameters on electric discharge machined
The corrosion potential of the EDM machined surface (Cu enriched) job surface integrity. Am J Eng Appl Sci 2010;3(2):396–402.
is greater (anodically polarized) compared to milled surface (bulk [13] ASM Handbook. Aluminum and aluminum alloys. In: Davis Joseph R., editor.
aluminum alloy). Thus, EDM surface will act as galvanic couple to Materials Park, OH: ASM International; 1993.
[14] Mayyas AT, Hamasha MM, Alrashdan A, Hassan AM, Hayajneh MT. Effect of copper
an adjacent bulk surface. and silicon carbide content on the corrosion resistance of Al-Mg alloys in acidic and
• Corrosion rates of the EDM machined aluminum surface were at alkaline solutions. J Minerals Mater Character Eng 2012;11(4):335–52.
[15] ASM Handbook. Properties and selection: nonferrous alloys and special purpose
least an order of magnitude higher than the milled surface.
•
materials, vol. 2; 1993.
Fatigue crack growth rates in a 0.06 M NaCl solution were increased [16] Porter DA, Eastling KE. Phase transformations in metals and alloys. London:
with an EDM notch compared to a cut notch for 7xxx and 2xxx series Champman & Hall; 1992.
alloys, 6xxx and 5xxx series alloys were not tested in this condition. [17] Davis JR. Corrosion of aluminum and aluminum alloys. Materials Park, OH: ASM
International; 1999.
[18] Blanc C, Lavelle B, Mankowski G. The role of precipitates enriched with copper on
The present study indicates that the EDM process should be used the susceptibility to pitting corrosion of the 2024 aluminum alloy. Corros Sci
with caution when manufacturing environmental fatigue test coupons. 1997;39:495–510.
[19] Bohni H, Suter T, Schreyer A. Micro-and nanotechniques to study localized corro-
However, there are remaining questions that could be addressed by sion. Electrochim Acta 1995;40(10):1361–8.
future researchers including, can EDM parameters be adjusted to limit [20] Buchheit RG. A compilation of corrosion potentials reported for intermetallic
the amount of material deposited from the wire, could other wire ma- phases in aluminum alloys. J Electrochem Soc 1995;142(11):3994–6.
[21] Buchheit RG, Boger RK, Carroll MC, Leard RM, Paglia C, Searles JL. The electro-
terials be used that would limit the microstructural changes or to pas-
chemistry of intermetallic particles and localized corrosion in Al alloys. JOM
sivate the notch. These changes should be investigated in combination 2001;53(7):29–33.
was other surface finishing effects to determine if the EDM process zone [22] Burstein GT, Vines SP. Repetitive nucleation of corrosion pits on stainless steel and
could be completely removed. the effects of surface roughness. J Electrochem Soc 2001;148(12):B504–16.
[23] Frankel GS. Pitting corrosion of metals: a review of the critical factors. J
Electrochem Soc 1998;145(6):2186–98.
Acknowledgements [24] Galvele JR, De Micheli SMd. Mechanism of intergranular corrosion of Al-Cu alloys.
Corros Sci 1970;10(1):795–807.
[25] Leblanc P, Frankel GS. A study of corrosion and pitting initiation of AA2024-T3
This work was funded under a USAFA Broad Agency Announcement using atomic force microscopy. J Electrochem Soc 2002;149(6):B239–47.
(FA-7000-11-2-0011; FA-7000-14-2-0013) by the United States Office [26] Leclere TJR, Newman RC. Self-regulation of the cathodic reaction kinetics during
of the Secretary of Defense Corrosion Policy and Oversight Office under corrosion of AlCu alloys. J Electrochem Soc 2002;149(2):B52–6.
[27] Ramgopal T, Gouma PI, Frankel GS. Role of grain-boundary precipitates and solute-
the Technical Corrosion Collaboration (TCC). depleted zone on the intergranular corrosion of aluminum alloy 7150. Corrosion
The views and conclusions contained herein are those of the authors 58(8); 2002: 687–97.
and should not be interpreted as necessarily representing the official [28] Scully JR, Peebles DE, Romig Jr AD, Frear DR, Hills CR. Metallurgical factors in-
fluencing the corrosion of aluminum, Al-Cu, and Al-Si sputtered thin films in dilute
polices and endorsements, either expressed or implied of US Air Force
HF solution. Metall Trans A 1992;23(9):2641–55.
Academy or the US Government. This material is declared a work of the [29] Suter T, Alkire RC. Microelectrochemical studies of pit initiation at single inclusions
United States Government and is not subject to copyright protection in in Al 2024–T3. J Electrochem Soc 2001;148(1):B36–42.
[30] Wall FD, Martinez MA. A statistics-based approach to studying aluminum pit in-
the United States.
itiation intrinsic and defect-driven pit initiation phenomena. J Electrochem Soc
2003;150(4):B146–57.
References [31] Warner JS, Gangloff RP. Alloy induced inhibition of fatigue crack growth in age-
hardenable Al–Cu alloys. Int J Fatigue 2012;42:35–44.
[32] Galyon Dorman SE, Rausch JW, Arunachalam SR, Burns JT, Gangloff RP, Fawaz SA,
[1] Field M, Kahles JF. The surface integrity of machined and ground high strength et al. Managing environmental impacts of time-cycle dependent structural integrity
steels. DMIC Rep 1964;210:54–77. of high performance DoD alloys. SAFE Inc., SAFE-RTP-16-045; 2016.
[2] Saoubi RM, Outerio JC, Chanderasekarn H, Dillon Jr. OW, Jawahir IS. A review of [33] Gangloff RP. Environment induced cracking of metals. In: Gangloff RP, Ives MB,
surface integrity in machining and its impact on functional performance and life editors. NACE; 1989. p. 55–109.
prediction of machined products. Int J Sustain Manuf 2008;1(1/2):203–36. [34] Wei RP, Gangloff RP. Fracture mechanics: perspectives and directions, ASTM STP
[3] Ultan D, Ozel T. Machining induced surface integrity in titanium and nickel alloys: a 1020. In: Wei RP, Gangloff RP. ASTM International; 1989. p. 233–64.
review. Int J Mach Tools Manuf 2011;51(3):250–80. [35] Ford FP, Andresen PL. Corrosion in nuclear systems: environmentally assisted
[4] Selvakumar G, Sornalatha G, Sarkar S, Mitra S. Experimental investigation and cracking in light water reactors. Corros Technol-New York and Basel
multi-objective optimization of wire electrical discharge machining (WEDM) of 2002;17:605–42.
5083 aluminum alloy. J Trans Nonferrous Metals Soc China 2014;24I(ssue2):373–9. [36] Ford FP, Andresen PL. Advances in fracture research. In: Salema K, editor.
[5] Dave HK, Mathai VJ, Desai KP, Raval HK. Studies on quality of microholes gener- Pergamon Press; 1989. p. 1571–84.
ated on Al 1100 using micro-electrodischarge machining process. Int J Adv Manuf
Technol 2013;76(1–4):127–40.

53