Electrochemical characteristics of intermetallic phases in aluminum alloys : An experimental survey and discussion

Article abstract of DOI:10.1149/1.1869984

This paper presents a survey of corrosion potentials, pitting potentials, and electrochemical characteristics for intermetallic particles commonly present in high-strength aluminum-based alloys. Results from relevant pure metals and solid solutions are also presented. It is seen that corrosion potentials and pitting potentials vary over a wide range for various intermetallics. Elaboration of the results reveals that the electrochemical behavior of intermetallics is more detailed than the simple noble or active classification based upon corrosion potential or estimated from the intermetallic composition. Intermetallics capable of sustaining the largest cathodic current densities are not necessarily those with the most noble E_corr, similarly those with the least noble E_corr will not necessarily sustain the largest anodic currents. The data herein was collected via the use of a microcapillary electrochemical cell facilitating electrode investigations upon intermetallic particles in the micrometer-squared range. This survey may be used as a tool for clarification of localized corrosion phenomena in Al alloys.

Full text of DOI:10.1149/1.1869984

B140
Journal of The Electrochemical Society, 152 ͑4͒ B140-B151 ͑2005͒
0013-4651/2005/152͑4͒/B140/12/$7.00 © The Electrochemical Society, Inc.
Electrochemical Characteristics of Intermetallic Phases
in Aluminum Alloys
An Experimental Survey and Discussion
,z
*
*
N. Birbilis and R. G. Buchheit
Fontana Corrosion Center, Department of Materials Science and Engineering, The Ohio State University,
Columbus, Ohio 43210, USA
This paper presents a survey of corrosion potentials, pitting potentials, and electrochemical characteristics for intermetallic
particles commonly present in high-strength aluminum-based alloys. Results from relevant pure metals and solid solutions are also
presented. It is seen that corrosion potentials and pitting potentials vary over a wide range for various intermetallics. Elaboration
of the results reveals that the electrochemical behavior of intermetallics is more detailed than the simple noble or active classifi-
cation based upon corrosion potential or estimated from the intermetallic composition. Intermetallics capable of sustaining the
largest cathodic current densities are not necessarily those with the most noble Ecorr, similarly those with the least noble Ecorr will
not necessarily sustain the largest anodic currents. The data herein was collected via the use of a microcapillary electrochemical
cell facilitating electrode investigations upon intermetallic particles in the micrometer-squared range. This survey may be used as
a tool for clarification of localized corrosion phenomena in Al alloys.
© 2005 The Electrochemical Society. ͓DOI: 10.1149/1.1869984͔ All rights reserved.
Manuscript submitted September 1, 2004; revised manuscript received November 12, 2004.
Available electronically March 14, 2005.
Background.—The desirable mechanical properties of many
self is a matter of some importance for emerging damage accumu-
lation models. For these models to be predictive, it is necessary to
develop a comprehensive, self-consistent accounting of this type of
pitting. In cases where the electrochemical characteristics of con-
stituent particles have been rigorously characterized, they have been
found to be much richer and more complicated than simple first-
order characterizations like noble or active.14 The aim of this work
is to develop a detailed characterization of the electrochemical be-
havior of intermetallics found in AA7075, many of which are appli-
cable to several other alloys.
commercial aluminum alloys are developed as a result of heteroge-
neous microstructures, stimulated by careful alloying additions and
heat-treatments. This is especially true in high strength Al alloys for
aerospace applications.1 From a localized corrosion perspective, the
dominant feature of alloy microstructures is the distribution of
second-phase ͑intermetallic͒ particles.2 Commonly such particles
will exhibit electrochemical characteristics that differ from the be-
havior of the matrix, rendering the alloy susceptible to localized
forms of corrosion. Over the years, a number of studies have been
carried out in order to assess the effect of specific intermetallic
particles and individual alloying additions upon corrosion accumu-
lation in Al alloys,3-13 which initiates from pitting-type corrosion.
Pit initiation is generally believed to begin by the rupture or
breakdown of passive film upon the metal surface.14 In the case of
pure metals, pitting resistance is dependant upon the electrochemical
stability of the passive film. However in the case of Al alloys, pitting
is influenced by the intermetallic particles which exhibit different
surface film characteristics to the matrix,15 and may be either anodic
or cathodic relative to the matrix.8 In Al alloys, two main types of
pit morphologies are observed.16-18 So called circumferential pits
appear as a ring of attack around a more or less intact particle or
particle colony. The attack appears to be mainly in the matrix phase.
This type of morphology has been ascribed to localized galvanic
attack of the more active matrix by the more noble particle. Leclère
and Newman19 offered another interpretation based on local pH gra-
dients, consistent with the concept of cathodic corrosion.
Microstructure considered.—The microstructures developed in
high-strength aluminum alloys such as AA7075, are complex and
incorporate a combination of equilibrium and nonequilibrium
phases. Typically such alloys have a chemical composition incorpo-
rating up to ten alloying elements. Such elements primarily include
Zn, Mg, and Cu, however appreciable and specific amounts of Fe,
Si, Cr, Ti, Zr, and Mn are often present ͑both as deliberate additions
and as impurities͒. Although presently not fully understood in terms
of its precise evolution, the microstructure of AA7075 is well
characterized,1,21 along with the corresponding physical metal-
lurgy.23 The literature quotes evidence supporting the presence of
the following intermetallics in 7XXX series alloys ͑not all are si-
multaneously present, and temper and precise composition will
regulate the types and proportion͒; Mg2Si, MgZn2, Al20Cu2Mn3,
Al12Mn3Si, Al7Cu2Fe, Al2Cu, Al2CuMg, Al3Fe, Al12Mg2Cr,
Al20Cu2Mn3, Al6Mn, Al3Ti, Al6Zr, Mg2Al3, Al32Zn49, and
Mg͑AlCu͒.1,22,24-27 The role of these intermetallics with respect to
mechanical properties is beyond the scope of this paper; however
intermetallic particles in aluminum alloys can be classified into three
main types.1
Precipitates form by nucleation and growth from a supersaturated
solid solution during natural or low-temperature artificial aging.
They range in size from Angstroms to fractions of a micrometer.
They can be spherical, needles, laths, plates, among other shapes.23
When they are homogeneously dispersed, their effect on localized
corrosion behavior is difficult to discern. However, when they are
concentrated on grain boundaries, they may affect intergranular cor-
rosion and stress corrosion cracking susceptibility.11 Important al-
loying elements leading to precipitation in Al alloys include Cu, Mg,
Si, Zn, Li.1 Typical examples include, MgZn2, Mg2Al3, Al2Cu, and
Al32Zn49.
The second morphology is apparent selective dissolution of the
constituent particle. Pits of this type are often deep and may have
remnants of the particle in them.16 This morphology has been inter-
preted as particle fall-out, selective particle dissolution in the case of
electrochemically active particles, or in the case of some Cu-bearing
particles, particle dealloying and non-faradaic liberation of the Cu
component.20
The different variety of particles that may appear in Al alloys is
vast,1,21-23 however those which may appear in 7XXX series alloys21
are of great technical interest and predominantly dealt with in this
paper. In order to understand the electrochemical behavior of such
classes of alloy, a comprehensive study on the electrochemical be-
havior for the range of intermetallic particles which can be present,
is warranted. How constituent particle-induced pitting manifests it-
Constituent particles are comparatively large and irregularly
shaped. Characteristic particle dimensions range from a few tenths
of a micrometer up to 10 micrometers. These particles can be
*
Electrochemical Society Active Member.
z E-mail: birbilis.1@osu.edu
Downloaded on 2014-12-08 to IP 130.194.20.173 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 152 ͑4͒ B140-B151 ͑2005͒
B141
formed during alloy solidification and are not appreciably dissolved
during subsequent thermomechanical processing. Rolling and extru-
sion tends to break up and align constituent particles into bands
within the alloy. Often constituents are found in colonies made up of
several intermetallic crystals or several different compound types.
Because these particles are rich in alloying elements, their electro-
chemical behavior is often significantly different than the surround-
ing matrix phase.2 In high strength Al alloys pitting is nominally
associated with some fraction of the constituent particles present in
the alloy.16-18 A range of alloying elements are found in constituents,
but Cu, Fe, Si, Mn, Mg are found most commonly.1 Typical ex-
amples include Al3Fe and Al7Cu2Fe.
Dispersoids are small particles comprising alloying elements that
are highly insoluble in aluminum. Particle sizes typically range from
0.05 to 0.5 ␮m. Cr, Ti, Zr, and Mn are common dispersoid formers.
With the possible exception of Mn the alloying elements concen-
trated in dispersoids are quite passive in environments where Al
alloys are regularly used. These particles form at high temperatures
and are present to control grain size and recrystallization behavior,
hence are homogeneously dispersed. Typical examples include
Al3Ti, Al6Mn, Al20Cu2Mn3, and Al3Zr.
formation regarding the electrochemical characteristics of interme-
tallic particles. Andreatta and co-workers42 have recently employed
the use of an electrochemical microcell ͑which they denote as a
microcapillary cell͒ for obtaining localized polarization data upon
AA7075. Although this technique may significantly reduce the area
of the working electrode, the reductions are not significant enough
to isolate individual particles or even particle-matrix interfaces.
Consequently, the results reflect testing upon a heterogeneous elec-
trode surface and again do not provide specific information regard-
ing intermetallic electrochemistry.
One major limitation of most of the studies discussed above, is
that the discussion of intermetallic particles is, often necessarily,
based upon qualitative assessments of intermetallic identification.
Although most often analyzed by energy dispersive X-ray spectros-
copy ͑EDXS͒, the EDXS method is not entirely accurate ͑indepen-
dently͒ in the identification of intermetallics, since X-rays may and,
most likely will, come from deeper within the sample than the in-
termetallic extends. This has led some authors to identify interme-
tallics more broadly, for example Al-Mg particles.14,37
Microelectrochemical testing.—In this study, we have employed
the synthesis of intermetallic analogs by specific casting. Following
this, electrochemical characterization was carried out via the use of
a microcapillary electrochemical cell42-46 ͑denoted in this work sim-
ply as microcell͒ capable of performing electrochemical testing on
electrode surfaces in the micrometer squared range.
Electrochemical testing using microelectrodes has become rather
widespread over the past decade,24,32,43-46 owing to the simplicity of
the methodology involved. In essence, microelectrochemical testing
may be considered identical to regular ͑bulk͒ electrochemical test-
ing, except that the size of the working electrode is reduced. This is
achieved by defining a working electrode as the area of metal that
comes into contact with a carefully placed solution, which either
takes the form of liquid droplet head, or the solution contained with
a capillary opening in the micrometer squared range.45 The critical
requirement of microelectrochemical testing is that the potentiostat
being used is capable of measuring low currents ͑typically
Ͻ10−10 A͒, since the working electrode area, hence associated cur-
rent, will be very small.
In regards to localized corrosion, the intermetallics of particular
interest are those which appear in the greatest proportion, either
by size or by frequency. For AA7X75 such particles have been
identified ͑in random order͒ as Mg2Si, MgZn2, Al7Cu2Fe, Al2CuMg,
Al2Cu, and Al3Fe25,27-29 ͑however again this can vary depen-
ding upon individual temper24͒. The matrix phase of AA7X75
is nominally comprised of Al-͑3-4 wt %͒Zn-͑2-3 wt %͒
Mg-͑0.5-1 wt %͒Cu.
Electrochemical testing of intermetallic compounds.—Electro-
chemical testing of intermetallic compounds synthesized in bulk
form has previously been used to study localized corrosion
phenomena,3,4,6,7,10,11,30-35 namely the role of microgalvanic interac-
tions in corrosion of Al-alloys.32 This approach has led to develop-
ments in the understanding of corrosion accumulation, for example,
it has been noted that the intermetallics Al2Cu and Al7Cu2Fe are less
prone to attack than Al2CuMg ͑S-phase͒.32 Alternatively, Liao and
Wei36 carried out studies using a zero resistance ammeter to galvani-
cally couple Al to two separate model alloys ͑an Al-Fe alloy and an
Al-Fe-Cu-Mn alloy͒. They found significant galvanic currents be-
tween Al and the model alloys, leading them to suggest galvanic
coupling can lead to significant corrosion for the couples investi-
gated. Scully and co-workers have also investigated the effect of
chromate upon several Cu-bearing intermetallics,12 whereas New-
man and co-workers have studied the effect of alloying elements in
Al solid solutions.3,19
In general, however, technical feasibilities mainly associated
with the small size of intermetallic particles, and the ability to per-
form electrochemical characterization on the microscale have not
yet allowed for several of the commonly observed intermetallics to
be studied. Furthermore, the polarization behavior which contains
more detailed electrochemical information as opposed to the corro-
sion potential alone, has not yet been characterized for a significant
number of the commonly occurring intermetallics in Al alloys.
The last few years have seen significant activity in the study of
microstructure-corrosion relationships for high strength Al alloys
͑namely AA2024 and AA7075͒.16,24,37-41 These studies have in-
volved the study of bulk alloy samples, with the techniques used to
assess corrosion capable of probing information on the micro- and
nanoscales. For example, Frankel and co-workers,37 and Andreatta
et al.38,39 have used scanning Kelvin probe force microscopy experi-
ments to assess local differences in the Volta potential of the alloy
surface at various stages of the corrosion cycle. Information regard-
ing corrosion damage morphology has also been gathered by use of
fluorescence microscopy16 and confocal scanning laser
microscopy.40 The abovementioned studies have added immensely
to the understanding of the extent and morphology of damage accu-
mulation upon such alloys, but have not provided fundamental in-
Experimental
Intermetallic synthesis.—Intermetallic particles were synthe-
sized by combining the necessary proportions of constituent pure
metals ͑supplied by Alfa-Aeser͒. Satisfactory synthesis of such par-
ticles requires prior knowledge of equilibrium conditions under
which the intermetallic in question will form. As a result, an equi-
librium phase diagram including the intermetallic required was
sourced and the composition required to fall within the appropriate
phase field was used. In many cases, intermetallics do not appear in
a unique phase field, and will appear in conjunction with other
phases.21 Consequently the feedstock proportions may not be indica-
tive of the stoichiometry of the intermetallic, but indicative of the
phase field which includes the given intermetallic. The appropriate
phase diagrams for phases present in Al alloys are readily available
throughout the physical metallurgy literature.21
Melting was carried out via two methods; either by induction
melting, or by arc melting. In all cases an argon atmosphere and
slight vacuum conditions were employed, as a safeguard to avoid
oxidation and vaporization, in order to produce laboratory ingots in
the vicinity of 150 g. Appropriate isothermal heat-treatments were
then employed to stimulate growth of the desired intermetallics
within the ingot or to generate the appropriate solid solution ͑in both
cases to comply with the appropriate equilibrium phase diagram͒,
with subsequent water quenching then carried out. In some cases,
the intermetallic compounds were sourced from previous studies.
The source and/or processing route for intermetallic synthesis is
included in Table I. Ingots incorporating Cr-containing intermetal-
lics, such as Al12Mg2Cr, could not be synthesized such that inter-
metallic particles would be large enough for microcell testing ͑hence
Downloaded on 2014-12-08 to IP 130.194.20.173 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

B142
Journal of The Electrochemical Society, 152 ͑4͒ B140-B151 ͑2005͒
Table I. Physical information related to intermetallic particles used in this work.
they are not discussed any further͒. A requirement of subsequent-
electrochemical testing via the microcell method dictated by capil-
lary tip diameter within our laboratory, was that intermetallic par-
ticles were larger than 50 ␮m in each of their surface dimensions.
Consequently, provided a small number of such particles were pro-
duced, then replicated electrochemical testing could be done.
It should be emphasized that the synthesis of homogeneous bulk
͑multigrained͒ intermetallic ingots is not readily possible. This is
owing to the fact that most intermetallics will not develop into
single-phase ingots under equilibrium conditions. Consequently, it is
not our methodology to create bulk ͑defect-free͒ single-phase inter-
metallic ingots. Instead, we are aiming to develop an ingot that will
include a population of the intermetallic crystals in question, in a
size range suitable for microcell testing. This therefore takes the
emphasis away from ingot production, and transfers it towards in-
termetallic characterization within the ingot. We do not advocate the
general use of bulk intermetallic analogs since it is difficult to en-
sure that the ingot is indeed homogeneous and indicative of the
intermetallic ͑furthermore bulk ingots will include grain boundaries
which may differ significantly in composition with respect to the
bulk ingot͒.
In order to evaluate the electrochemical behavior of the matrix
phase of AA7075, it was decided to use a particle-free matrix
equivalent. As a result, the matrix phase of AA7475 was deemed to
serve as an appropriate equivalent, owing to the fact that the alloys
have similar amounts of Zn, Mg, and Cu ͑hence forming a similar
solid solution͒, however AA7475 has very low alloy values for Fe,
Si, Zr, Ti, and Mn,1 largely avoiding the formation of inclusions,
dispersoids, and submicrometer-sized particles, will largely be
avoided. Another factor is the lack of particles which may act as
grain nucleators or grain refiners, resulting in a slightly larger grain
size than that of AA7075, hence allowing the matrix phase of
AA7475 to serve as a suitable particle-free and large-phase which
may be used to tentatively represent the electrochemical behavior of
particle free AA7075.
Downloaded on 2014-12-08 to IP 130.194.20.173 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 152 ͑4͒ B140-B151 ͑2005͒
Table I. (Continued.)
B143
Intermetallic identification.—Satisfactory identification of inter-
metallics requires both chemical and structural information. Charac-
terization of the intermetallics was done via a scanning electron
microscopy ͑SEM, Philips XL-30 FEG-ESEM͒ enabled with quan-
titative energy dispersive X-ray spectroscopy ͑EDXS͒, and the fa-
cility to generate backscattered electron Kikuchi patterns ͑BEKPs͒.
The EDXS was capable of providing details of the composition of
the crystals being probed ͑Table I͒. In order to translate this into
accurate intermetallic identification, the supplementary use of
BEKPs was employed, as previously done in Ref. 20. Following
application of an intense electron beam ͑supplied by a field emission
gun͒, spatial resolution of intercepted backscattered electrons yield-
ing structural information is facilitated by a carefully placed camera
detector. This detector is different from that responsible for back-
scattered electron imaging. Nominally the sample is tilted in the
SEM chamber ͑commonly to 70°͒ such that the sample surface is
perpendicular to the detector responsible for generation of Kikuchi
patterns.
A typical example of a BEKP for Al6Mn is included in Fig. 1
along with the corresponding EDXS spectrum. Each intermetallic
synthesized in this work has a unique crystal structure as previously
determined.21 Table I shows the simulated pattern overlays on the
collected BEKPs for all synthesized samples, indicating agreement
between the collected and simulated patterns. The BEKP collection
and analysis was facilitated by OIM software ͑TSL®͒. Sample
preparation for the generation of satisfactory BEKPs requires that
the sample surface be meticulously flat ͑i.e., polished to below
0.1 ␮m finish͒. In this work samples were prepared by metallo-
1
graphic polishing down to ␮m finish, and subsequently vibratory
polished using a 0.05 ␮m g4amma alumina suspension. This suspen-
Downloaded on 2014-12-08 to IP 130.194.20.173 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

B144
Journal of The Electrochemical Society, 152 ͑4͒ B140-B151 ͑2005͒
sion was chosen owing its inertness, in an effort to avoid surface
dissolution.
Electrochemical characterization.—Electrochemical testing of
the synthesized intermetallics was carried out using a microcell
method, as outlined previously in Ref. 32 and 43-45. In this method,
the working electrode area is defined by the area of metal which
comes into contact with the opening of a microcapillary. The micro-
capillary is filled with electrolyte/solution, while containing a small
wire counter electrode and electrolytic contact with a saturated
calomel reference electrode. A silicone skirt/seal was applied to the
open end of the capillary in order to avoid any solution leakage and
to allow an interference contact with the working electrode. The
capillary opening is generally in the vicinity of 20-60 ␮m diam, and
will vary with each capillary. Therefore, each test must be treated
individually to establish the correct working electrode area for de-
termination of current density values. A series of schematics and
pictures is included as Fig. 2, showing various aspects of the micro-
cell method as adopted in this work. Figure 2a indicates a schematic
of the electrochemical cell component used to carry out testing. A
more complete set of pictures covering apparatus and setup can be
found in Ref. 44. Figure 2b indicates a typical microcapillary used,
with the silicone skirt clearly visible. As an example to the spatial
resolution of the method, a picture of an Al20Cu2Mn2 crystal is
shown prior to electrochemical testing and also following testing
͑Fig. 2c and d͒. The microcell used in these studies was incorporated
into a lenspiece of an optical microscope. This made it possible to
switch between viewing/aligning the sample on the microscale, and
subsequent testing with the microcell. The microscope was also con-
nected to a video output, allowing microimages to be captured digi-
tally.
Figure 1. Backscattered electron Kikuchi pattern and corresponding EDXS
spectrum for Al6Mn crystal.
Potentiodynamic polarization was carried out at a scan rate of
Figure 2. ͑a͒ Schematic representation of
microcapillary electrochemical cell. ͑b͒
Glass capillary with tip coated in silicone.
͑c͒ Optical micrograph of Al20Cu2Mn3
crystal prior to testing. ͑d͒ Optical micro-
graph of Al20Cu2Mn3 crystal following
electrochemical testing with the microcell.
Test area is evident.
Downloaded on 2014-12-08 to IP 130.194.20.173 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 152 ͑4͒ B140-B151 ͑2005͒
B145
Figure 4. Typical polarization curves for aluminum collected using the mi-
crocell method at three different chloride concentrations at neutral pH. The
values for Ecorr and Epit are noted.
Figure 3. Pitting potentials for aluminum presented as a function of chloride
ion activity.
0.01 V s−1 using an Autolab PGSTAT 100 with resolution at low
currents ͑down to 1 ϫ 10−15 A͒. Equilibration time for the working
electrode was 60 s following contact with the solution. Such precau-
tions were taken to minimize test durations, in order to avoid any
issues with cell leakage, or the establishment of concentration gra-
dients in the capillary. It should be pointed out that the nominal cell
resistance ͑responsible for iR-drop͒ was comparatively high with
respect to other systems, with values approaching 10 k⍀ recorded.
However, since the absolute currents measured during testing rarely
approached values up to 10−7 A ͑owing to the small size of the
working electrode͒, ohmic drop was considered negligible for po-
tentiodynamic testing, since it would be nominally confined to under
1 mV. The tests reported herein should also be considered to have
been carried out in aerated conditions, since no special effort was
made to exclude oxygen from the capillary or solution. Each test
was repeated at least six times and the average values are reported
herein. Measurements were performed in 0.01, 0.1, and 0.6 M NaCl
at pH 6.
evant pure metals, solid solutions, and an analog to the matrix phase
of AA7075.
Table II. Corrosion potentials for intermetallic compounds com-
mon in aluminum alloys. Results are average values incorporat-
ing the results of numerous tests.
Corrosion potential ͑mVSCE
͒
Stoichiometry
Phase
0.01 M
0.1 M
0.6 M
Note
Al3Fe
Al2Cu
Al3Zr
Al6Mn
Al3Ti
Al32Zn49
Mg2Al3
MgZn2
Mg2Si
Al7Cu2Fe
Mg͑AlCu͒
Al2CuMg
Al20Cu2Mn3
Al12Mn3Si
Al ͑99.9999͒
Cu ͑99.9͒
Si ͑99.9995͒
Mg ͑99.9͒
Mn ͑99.9͒
Cr ͑99.0͒
Zn ͑99.99͒
Al-2%Cu
Al-4%Cu
7X75 Matrix
AA 7075-T651
␤
␪
␤
-
␤
T
Ј
␤
M, ␩
␤
-
-
S
-
-
-
-
␣
-
-
−493
−592
−752
−839
−620
−1009
−1124
−1001
−1355
−549
−898
−956
−550
−890
−679
−177
−450
−1601
−1315
−495
−985
−813
−750
−699
−816
−539
−665
−776
−779
−603
−1004
−1013
−1029
−1538
−551
−943
−883
−565
−810
−823
−232
−441
−1586
−1323
−506
−1000
−672
−602
−799
−965
−566
−695
−801
−913
−799
−1063
−1162
−1095
−1536
−654
−936
−1061
−617
−858
−849
−220
−452
−1688
−1318
−571
−1028
−744
−642
−812
−1180
Results
Verification of microcell data.—Given the relatively unique na-
ture of electrochemical testing via the microcell method, owing to
the rapid scan rates used and the short equilibration times prior to
polarization, a set of control tests was initially conducted upon pure
Al ͑99.9999͒. This was done in order to correlate with previously
reported literature values and to verify the sensitivity of the method
to variations in electrochemical response with respect to electrolyte
aggressiveness. The results of breakdown/pitting potential ͑Epit͒ vs.
chloride activity are given in Fig. 3, and compared with those of
McCafferty,47 Kaesche, Bohni, and Uhlig ͑given in Ref. 48͒ who
performed tests upon AA 1199 ͑Al 99.99͒. Typical potentiodynamic
polarization curves that went into the construction of Fig. 3 are
included in Fig. 4. The raw data in Fig. 4 was collected using the
microcell method. Based upon the sensitivity of the microcell
method as evidenced by the control testing upon aluminum ͑Fig. 3
and 4͒, it is posited the method is able to delineate between the
electrochemical behaviors of various intermetallic particles, in spite
of the relatively unique experimental parameters employed by the
technique. As a result, meaningful potentiodynamic diagrams may
be obtained.
A
A
A
A
A
A
A
B
-
-
␣
␣
-
B
M
X
-
Notes:
A. Pure metal were obtained from Alfa-Aeser and tested using the micro-
cell method.
B. These specimens are homogeneous solid solutions and tested using the
microcell method.
M. The phase denoted as 7X75 matrix is the particle-free matrix-phase of
AA7474.
Electrochemical behavior of intermetallic particles.—The cor-
rosion potentials ͑Ecorr͒ for the compounds tested in this study are
given in Table II. Table II also includes information relating to rel-
X. Tests upon AA7075-T651 were done on bulk specimens using conven-
tional electrochemical methods and an electrode area of 1 cm2.
Downloaded on 2014-12-08 to IP 130.194.20.173 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

B146
Journal of The Electrochemical Society, 152 ͑4͒ B140-B151 ͑2005͒
Table III. Pitting potentials for intermetallic compounds common
in aluminum alloys. Results are average values incorporating the
results of numerous tests.
Pitting potential ͑mVSCE
͒
Stoichiometry
Phase
0.01 M
0.1 M
0.6 M
Note
Al3Fe
Al2Cu
Al3Zr
Al6Mn
Al3Ti
Al32Zn49
Mg2Al3
MgZn2
Mg2Si
␤
␪
␤
-
442
−434
−223
−485
−232
-
−818
-
-
106
−544
−275
−755
−225
-
−846
-
-
−382
−652
−346
−778
−646
-
−959
-
-
␤
T
C
Ј
␤
M, ␩
␤
C
C
Al7Cu2Fe
Mg ͑AlCu͒
Al2CuMg
Al20Cu2Mn3
Al12Mn3Si
Al ͑99.9999͒
Cu ͑99.9͒
Si ͑99.9995͒
Mg ͑99.9͒
Mn ͑99.9͒
Cr ͑99.0͒
Zn ͑99.99͒
Al-2%Cu
Al-4%Cu
7X75 Matrix
AA 7075-T651
-
-
S
-
-
-
-
␣
-
-
-
-
␣
␣
-
−447
224
108
−210
−563
−545
19
−448
−2
80
−428
−621
−610
−30
-
−1391
-
297
-
−471
−406
−736
−739
−580
-
D, E
F
135
−534
−712
−696
−94
-
−1473
-
190
-
A
A
A, C
A, G
A, C
A
-
−1095
-
479
-
−447
−418
−633
−684
A, C
B
B
Figure 5. Corrosion potential ͑Ecorr͒ vs. NaCl concentration ͑of test solution͒
for various intermetallic compounds and pure Al ͑99.9999͒.
−529
−465
−768
−810
M
X
-
from Table II ͑since the data in Table II gives the average values
only͒. In order to visualize this, Ecorr results for a number of inter-
metallics have been summarized graphically in Fig. 5, along with
pure Al. Not all compounds have been plotted in order to allow a
feasible number of points to be shown on one graph.
Notes:
C. These compounds do not shows a breakdown of passivity, with active
dissolution occurring at potentials more positive than Ecorr
D. Did not show a breakdown in all cases when tested at 0.1 M NaCl.
E. At the highest concentration of NaCl tested, this compound did not
display a breakdown of passivity, with active dissolution occurring at poten-
.
Dispersion in characteristic potentials.—From Fig. 5 it is appar-
ent that the measured Ecorr values vary within a range represented by
the error bars between replicated tests. Typically, Ecorr values varied
within a window of about 80 mV between replicate tests, whereas
Epit values typically varied within about a 40 mV window. Data
distribution between repeated tests is common,34 and Figs. 6a-e
show the distributions corresponding to the intermetallics Al2Cu,
Al7Cu2Fe, MgZn2, Mg2Si, and Al2CuMg to serve as general ex-
amples. We note that there is dispersion in the results presented in
Figs. 6a-e. However, it is unlikely that this dispersion will compro-
mise the overall validity of the electrochemical testing herein, but
merely serves to highlight that characteristic electrochemical param-
eters do vary within an envelope.
tials more positive than Ecorr
.
F. The breakdown potential ͑Epit͒ of S-phase should be viewed with cau-
tion. The electrochemical behavior of this compound is complex20 and incor-
porates dealloying of the Al and Mg, capable of generating a relatively large
corrosion current density prior to ultimate breakdown. For more details re-
garding S-phase, see Ref. 20 and 30.
G. The quoted Epit values of pure Mg correspond with the potential at
which current density rapidly increases. Pure Mg, however, is generally un-
stable and freely corrodes in Cl-containing solution.
The corresponding pitting potentials ͑Epit͒ for the compounds
tested in this study are given in Table III. We note that not all
compounds tested show a characteristic breakdown potential. In
such cases, corrosion current gradually increases as potentials more
anodic than Ecorr are realized. Such compounds do not show any
Results viewed with relation to corrosion of AA7075.—A mon-
tage of polarization curves generated in 0.1 M NaCl for the matrix
analog and the corresponding intermetallics associated with AA7075
is presented in Fig. 7. This representation serves as a useful way in
which to view a series of polarization curves. For equivalent values
of potential, the individual behavior of the intermetallics vs. the
matrix may be compared. We note that at potentials corresponding
to the Ecorr of AA7075 ͑in NaCl solution͒, MgZn2 and Mg2Si are
experiencing extensive anodic dissolution. In contrast to this, at the
potentials corresponding to the individual Ecorr values of Al3Fe and
Al2Cu, the matrix is experiencing significant anodic dissolution. The
interpretation of Fig. 7, along with a discussion regarding the
S-phase, is expanded in the Discussion section.
passivity and hence corrode freely above Ecorr
.
The data provided in Tables II and III incorporates the results of
an extensive test program. As previously suggested, not all the com-
pounds listed in Tables II and III will play a key role in determining
the overall corrosion kinetics of AA7075. Those which are postu-
lated to be of principal significance are those which were previously
recognized as being present in the largest size and frequency. How-
ever, the complete list of compounds has been presented as it forms
a crucial part of this experimental survey, while the intermetallics
listed are relevant to a large number of commercial Al alloys.
Overall, a sense of the relative nobility for the compounds tested
may be gained by examination of Tables II and III: The trends in
results appear to be consistent irrespective of NaCl concentration.
The net effect of increasing NaCl concentration is that both the Ecorr
and Epit values are shifted to less noble values, similar to the manner
seen for pure Al.47 In some cases, this may not be entirely obvious
Additional results for intermetallics not extensively characterized
by electrochemical methods.—In order to provide additional kinetic
information regarding intermetallics not previously reported in ex-
tensive detail, the typical potentiodynamic polarization curves col-
lected in 0.1 M NaCl are shown for Al32Zn49, Mg͑AlCu͒, Al3Ti,
Downloaded on 2014-12-08 to IP 130.194.20.173 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 152 ͑4͒ B140-B151 ͑2005͒
B147
Figure 6. Distribution plots of Ecorr and
Epit for ͑a͒ Al2Cu, ͑b͒ Al7Cu2Fe, ͑c͒
MgZn2, and ͑d͒ Mg2Si, and ͑e͒ Al2CuMg
in 0.01 M NaCl solution.
ported by other authors and compiled by Buchheit ͑Ref. 2 and ref-
erences reported therein͒ or given in Ref. 11, 20, 31, 32, 34, 35, 48,
and 49. As noted in this compilation, information regarding electro-
chemical characteristics of intermetallic compounds is not available
in a unified publication; however this work has been able to draw
together a large amount of characteristic data in standard test solu-
tions to allow for a consolidated presentation. The magnitude and
depth of data in Tables II and III does not permit for a complete
discussion of all results in each context where they may apply, how-
ever in order to provide a working example we have focused on
corrosion of AA7075 as one example.
Mg2Al3, Al12Mn3Si, and Al7Cu2Fe ͑Fig. 8a-f͒. These results are
contrast with those already available in the literature in the follow-
ing section.
Discussion
Electrochemical behavior of intermetallic particles.—The tabu-
lated results indicate that the range of corrosion potentials observed
for the different intermetallics varies over several hundreds of mil-
livolts, reflecting the electrochemical heterogeneity that may be
found in many commercial Al alloys. We note that characteristic
values of Ecorr for intermetallics are dependant upon chemical com-
position. It is seen that compounds containing Cu, Fe, and Ti are
more noble than pure Al or the matrix analog. Such compounds
include Al3Fe, Al7Cu2Fe, Al2Cu, and Al3Ti. All of these compounds
reveal a characteristic breakdown potential, indicating that they are
capable of maintaining a passive film.
In contrast, intermetallics containing Mg, Zn, or Si, are typically
less noble than pure Al or the matrix analog. It is noted that
Al32Zn49, MgZn2, and Mg2Si do not show any breakdown potential
and are capable of corroding freely above their Ecorr. This is signifi-
cant, since the latter two are present to an appreciable extent in
several 7XXX series alloys.24,25,28-30
Discussion of results viewed in relation to 7075.—In the con-
text of AA7075, we can use the data herein to allow us to understand
the evolution of corrosion damage accumulation. For example, if we
recall the data presented in Fig. 7 ͑collected in 0.1 M͒, we can make
certain specific assumptions regarding the corrosion morphology we
expect to see. In the case of MgZn2 and Mg2Si, we would anticipate
observing anodic dissolution of these intermetallic particles, ulti-
mately leaving behind a surface cavity ͑selective dissolution͒. The
reason for this is owing to the inability of these two intermetallics to
passivate; hence at potentials which may be realized when polarized
by/coupled to the matrix, dissolution will proceed until the interme-
The reported values of the corrosion potentials determined in this
study correlate with the applicable related values previously re-
Downloaded on 2014-12-08 to IP 130.194.20.173 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

B148
Journal of The Electrochemical Society, 152 ͑4͒ B140-B151 ͑2005͒
the underpinning of much of the mechanistic explanations in Al
alloy corrosion science.49 We now see an example of how little these
terms account for, since the behavior of intermetallics is seen as
being very rich. For example, Mg2Si displays a value of Ecorr several
hundreds of millivolts more negative than MgZn2; hence a tradi-
tional interpretation may suggest that this is the more active of the
two intermetallics. Figure 7, however, reveals that across a wide
range of potentials typical of the matrix, MgZn2 shows anodic dis-
solution about two orders of magnitude greater than Mg2Si. This
phenomenon is only detectible owing to the complete electrochemi-
cal characterization of these intermetallics, and is not obvious from
Ecorr measurements and estimates based on composition alone.
In contrast, for the case of Al2Cu and Al3Fe, we would expect to
see little or no dissolution of these noble particles. The ability of
such particles to efficiently sustain cathodic reactions has been pre-
viously noted.7 Hence, we may expect the matrix adjacent to the
particle to corrode owing to the polarization provided by the noble
particle and/or a localized increase in the pH ͑due to hydroxyl ion
formation͒ promoting dissolution of the adjacent matrix. Evidence
of so-called peripheral matrix dissolution is often observed.16
The richness in the behavior of intermetallic particles ͑IMPs͒ is
evident again, however, in the case of Al2Cu and Al3Fe. Since Al3Fe
displays an Ecorr value more noble than that of Al2Cu ͑also con-
comitant with a more noble Epit͒, one may expect that Al3Fe be
the most noble of the intermetallics. If this were the case, Al3Fe
should provide the largest driving force for the corrosion of the
adjacent matrix. What is seen, however, is that in a wide potential
range characteristic of the matrix, Al2Cu can sustain cathodic reduc-
tion processes at rates an order of magnitude higher than Al3Fe.
Thus, traditional concepts of relative activity based upon active or
noble are insufficient to describe mechanisms in overall damage
accumulation.
A special note must be made regarding the electrochemical be-
havior of the S-phase ͑Al2CuMg͒. The S-phase is unique since it
nominally displays an Ecorr value less noble than that of the matrix,
however, the high breakdown potential may lead one to believe that
the S-phase is cathodic to the matrix. The dissolution characteristics
of the S phase are complex and have been shown20 to incorporate a
dealloying mechanism, leading to localized Cu enrichment. As a
result, relatively large current densities prior to ultimate breakdown
may be seen, with the ultimate breakdown possibly corresponding to
that of nearly pure ͑redistributed͒ Cu. A detailed study regarding the
electrochemical characteristics of the S-phase is not given in this
paper, however more details may be found in Ref. 20 and 30. It may
be assumed that the S-phase does not dissolve as a unique entity, but
may undergo a dealloying process leading to selective dissolution of
the Mg and Al components of the intermetallic. The result of this
may be that the resultant Cu enrichment may allow for the interme-
tallic to behave as a local cathode after some unknown time.37
Elaboration of the above concepts is clearer by investigation of
Table IV. Table IV lists the compounds tested in galvanic series
order, based upon the evaluated Ecorr. In the specific case of
AA7075, we see that the majority of intermetallics are more noble
than the matrix. Those intermetallics which are deemed less noble
than the matrix are generally found to contain Mg and/or Zn.
Table IV also includes the average value of the corrosion current
measured at Ecorr during potentiodynamic testing. We nominally ob-
serve that the compounds showing the highest corrosion currents
are those more active ͑i.e., less noble͒ than the matrix. Such com-
pounds are likely to be dissolving themselves at a high rate within
the alloys.
Figure 7. Montage of polarization curves for phases typically present in
AA7075.
Information pertaining to the average currents measured upon
compounds tested at the potential corresponding to the Ecorr of the
AA7075 is also included in Table IV. This information provides a
deeper insight into the overall behavior of the alloy, further indicat-
ing the weaknesses of Ecorr values alone as a basis for mechanistic
interpretations of alloy corrosion, while revealing which IMPs may
warrant the most attention. It is seen that the largest currents at the
tallic is disintegrated. It may even be possible that some Mg enrich-
ment may be observed in the presence of alkaline solutions owing to
incongruent dissolution.28
Figure 7, however, allows us to challenge the generic concepts of
noble and active as applied to intermetallics, which presently form
Downloaded on 2014-12-08 to IP 130.194.20.173 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 152 ͑4͒ B140-B151 ͑2005͒
B149
Figure 8. Potentiodynamic polarization
curves for ͑a͒ Al32Zn49, ͑b͒ Mg͑AlCu͒, ͑c͒
Al3Ti, ͑d͒ Mg2Al3, ͑e͒ Al12Mn3Si, ͑f͒
Al17Cu2Fe, generated using the microcell
method and 0.1 M NaCl solution.
Ecorr of AA7075 can be sustained by MgZn2 ͑anodic͒, along with
Al7Cu2Fe and Al2Cu ͑cathodic͒. Hence these IMPs should form the
watch list in relation to corrosion kinetics in cases where they are
present. Whether these IMPs identified as the watch list are indeed
primarily responsible for overall corrosion kinetics is discussed in
the following section.
increase in the corresponding Ecorr, and a reduction in the dissolu-
tion kinetics of the intermetallic. The presence of Mg͑AlCu͒ has
been noted in high strength 7XXX series alloys.25
Al3Ti.—Al3Ti ͑Fig. 8c͒ shows a distinct passive-active transition
with Epit separated from Ecorr by about 400 mV. We also see classi-
cal positive hysteresis in the reverse scan. Al3Ti is one of the most
noble IMPs tested in this work, with an Ecorr more positive than
−0.6 VSCE. The values measured here correspond well with those of
Lavrenko et al.31 who determined a stationary corrosion potential of
−0.6 V in artificial seawater. Al3Ti is present in several high-
strength alloys, since Ti is intentionally added as an innoculant.
Consequently Al3Ti is very small and classed as a dispersoid. The
effect of Al3Ti on corrosion behavior of Al alloys has not previously
received significant attention, since the submicrometer size of Al3Ti
and the ability to resolve a corrosion-IMP relationship on the nanos-
cale is relatively difficult.
Electrochemical behavior of selected intermetallics not exten-
Al Zn .—The IMP Al Zn49 may appear in
sively characterized.—
32 49
32
certain 5XXX or 7XXX series alloys,23 and has been previously
discussed in the context of SCC of such alloys. It is seen that for
Al32Zn49 ͑Fig. 8a͒, there is no passive region evident. At potentials
more positive than Ecorr, dissolution rates increase markedly and
approach rates beyond 1 mA/cm2, even for relatively modest poten-
tials, i.e., at values in the vicinity of most Al alloys.
Mg(AlCu).—Similarly, there is no evident passive region for
Mg͑AlCu͒ ͑Fig. 8b͒, however, the corresponding dissolution rates
during anodic polarization remain comparatively lower than other
This IMP has recently received some detailed attention,
since it has been associated with the SCC behavior of 5XXX series
alloys. In spite of the relatively large amount of Mg present in this
Mg2Al3.—
nonpassivating compounds, even at potentials as high as +1 VSCE
.
Mg͑AlCu͒ can be considered analogous to MgZn2 with Cu and Al
substituting Zn sites. The effect of this substitution results in an
Downloaded on 2014-12-08 to IP 130.194.20.173 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

B150
Journal of The Electrochemical Society, 152 ͑4͒ B140-B151 ͑2005͒
Table IV. Galvanic series for compounds tested in this work.
Average free
corrosion rate at
corrosion potential
͑A/cm2͒
Average current at
corrosion potential
of AA7075-T651
͑A/cm2͒
Corrosion
potential
͑m VSCE͒
Stoichiometry
Phase
Mg ͑99.9͒
Mg2Si
MgZn2
-
␤
M, ␩
␤
−1586
−1538
−1029
−1013
−1004
−1000
−965
−943
−883
−823
−810
−799
−779
−776
−672
−665
−603
−602
−565
−551
−539
−232
5.5 ϫ 10−6
7.7 ϫ 10−6
8.4 ϫ 10−5
4.8 ϫ 10−6
1.4 ϫ 10−5
1.2 ϫ 10−6
1.07 ϫ 10−6
2.3 ϫ 10−5
2.0 ϫ 10−6
3.9 ϫ 10−6
1.7 ϫ 10−6
3.2 ϫ 10−6
6.3 ϫ 10−6
2.5 ϫ 10−6
1.3 ϫ 10−6
7.3 ϫ 10−6
5.6 ϫ 10−7
2.3 ϫ 10−6
3.4 ϫ 10−7
6.3 ϫ 10−6
2.1 ϫ 10−6
1.8 ϫ 10−6
1.5 ϫ 10−5
1.9 ϫ 10−4
1.0 ϫ 10−3
1.9 ϫ 10−5
2.9 ϫ 10−4
8.1 ϫ 10−5
-
Mg2Al3
Al32Zn49
Zn ͑99.99͒
AA 7075-T651
Mg ͑AlCu͒
Al2CuMg
Al ͑99.9999͒
Al12Mn3Si
7X75 Matrix
Al6Mn
T
Ј
-
-
-
S
-
-
−1.2 ϫ 10−5
−2.1 ϫ 10−6
−4.7 ϫ 10−5
−7.6 ϫ 10−5
−8.1 ϫ 10−5
−1.2 ϫ 10−4
−8.1 ϫ 10−5
−2.1 ϫ 10−4
−4.7 ϫ 10−4
−2.7 ϫ 10−4
−2.9 ϫ 10−4
−1.9 ϫ 10−5
−3.1 ϫ 10−4
−9.9 ϫ 10−5
−1.8 ϫ 10−3
-
-
Al3Zr
Al-2%Cu
Al2Cu
␤
␣
␪
␤
␣
-
-
␤
-
Al3Ti
Al-4%Cu
Al20Cu2Mn3
Al7Cu2Fe
Al3Fe
Cu ͑99.9͒
IMP, Mg2Al3 shows a distinct breakdown at values just below
−0.8 VSCE ͑Fig. 8d͒, indicating that it has a window of passivity in
the NaCl solution. This is supported by the subsequent positive hys-
teresis observed. It is interesting to note that breakdown occurs at a
Al7Cu2Fe.—The compound Al7Cu2Fe has recently been discussed
as a contributor to localized damage accumulation upon AA202412
and AA7075.40 This IMP shows a relatively noble Ecorr value ͑in the
vicinity of −0.6 VSCE͒ with respect to other intermetallics listed in
Table II. Concomitantly the Epit value, which is very distinct, occurs
at values of −0.45 VSCE, followed by positive hysteresis ͑Fig. 8f͒.
The effect of both Cu and Fe also allows for Al7Cu2Fe to sustain
relatively large cathodic current densities ͑Table IV͒, indicating that
it may be an archetypal noble particle with low dissolution rate and
a high efficiency for supporting oxygen reduction.
potential consistent with
a relatively large current density
͑0.1 mA/cm2͒. Furthermore the reversible potential on the reverse
scan shows a slight ennoblement to that of the forward scan.
Similar testing upon synthesized Mg2Al3 was published by
Searles and co-workers.11 Those results were obtained in buffered
solutions of aerated 3.5 pct NaCl solution, and it was seen that they
correspond well with those presented here. Across a wide range of
pH ͑3.3-13.5͒, Searles et al. quotes Ecorr values in the vicinity of
−1.25 VSCE, generally about 100 mV less negative than those seen
in Table II for our tests in 0.6 M NaCl. In regards to Epit, consistent
values in the vicinity of −0.95 VSCE are seen in Ref. 11 and in Table
III for tests in 0.6 M NaCl. Overall, Mg2Al3 may be considered
active, noting that the Epit value of Mg2Al3 are generally below the
corresponding corrosion potentials for alloys in which it may be
found.
General.—Although Table IV reveals that a large number of
intermetallics show electrochemical activity, not all compounds
listed may significantly impact corrosion kinetics. For example, as
previously mentioned, due to their small size, homogenous disper-
sion and comparative electrochemical inertness, dispersoids do not
appear to have a significant direct effect on localized corrosion sus-
ceptibility in high-strength aluminum alloys. A search of the litera-
ture does not reveal any evidence that Al20Cu2Mn3, Al3Zr, Al3Ti,
and Al12Mn3Si have impacted adversely on corrosion properties of
commercial Al alloys. Consequently in the interpretation and appli-
cation of the data included in this study, the effect of intermetallic
size should be considered ͑as this will govern the amount of current
the intermetallic can support͒. Intermetallics rich in Cu or Fe are
nominally orders of magnitude larger in size than dispersoids and
nominally capable of supporting oxygen reduction much more effi-
ciently.
Although not discussed in great detail, the supplementary data
included in Tables II and IV regarding ͑Al-Cu͒ solid solutions and
pure metals ͑Al, Mg, Mn, Zn, Si, and Cu͒ can help describe the
behavior of intermetallics and the role of pure elements in destabi-
lizing the passive film or enhancing oxygen reduction kinetics.
These results also form an important aspect of this survey for pro-
spective users of this data.
This IMP has previously been studied in the context of
Al12Mn3Si.—
3XXX series alloys and architectural alloys where filiform corrosion
may occur.10 Al12Mn3Si reveals a distinct breakdown, in this case at
−0.6 VSCE ͑Fig. 8e͒. This breakdown is associated with a slight
ennoblement in the reverse scan. One possible reason for why we
may observe such a slight ennoblement for the above two interme-
tallics may be due to selective dissolution of the more active com-
ponents, viz. Mg and Si, during the anodic cycle; however this has
not been independently confirmed in this work.
Although this IMP is generally classed as noble with respect to
the remainder of bulk alloys, we note that in the potential range
characteristic of many Al alloys, say between −0.8 to − 1 VSCE
,
that Al12Mn3Si does not show high rates of oxygen reduction. This
may suggest that this IMP does not play a significant role in the
overall corrosion of the alloys in which it is dispersed. This was
supported in a study by Afseth et al.10 where testing revealed that
Al12Mn3Si did not have any discernable impact on the corrosion of
AA3005.
The results presented herein for the Al-Cu solid solutions may be
discussed in accordance with the results of Leclère and Newman,19
where by the increase in Cu content allows for enhanced oxygen
reduction. This, however, raises an important point regarding inter-
Downloaded on 2014-12-08 to IP 130.194.20.173 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 152 ͑4͒ B140-B151 ͑2005͒
B151
2. R. G. Buchheit, J. Electrochem. Soc., 142, 3994 ͑1995͒.
metallic and alloy corrosion, in that overall more corrosion has been
previously observed upon samples richer in Cu. This is contradic-
tory to the simple noble and less noble concepts, which we are
challenging, and has been credited to local enrichment sites of Cu
which have the ability to function as pure cathodes ultimately lead-
ing to incongruent ͑self͒ dissolution. Such incongruent dissolution
has also been observed upon the S-phase,20 and may indeed form the
basis of self-corrosion for several of the intermetallics investigated
in this study. Consequently, further refinements to the classifications
herein based on surface analysis will allow us to account for
changes in electrochemical behavior with time or as a function of
environment due to the evolution of the surface film by incongruent
dissolution or dealloying.
Consequently we may form a tentative classification for interme-
tallics based on the work herein, which may be given as:
Noble particles with high electrochemical activity, e.g., Al2Cu,
Al7Cu2Fe. Ecorr Ͼ Ecorr of alloy, with the ability to sustain large
cathodic current. Such intermetallics are possibly associated to pe-
ripheral pitting.
Noble particles with low electrochemical activity, e.g., Al3Zr.
Ecorr Ͼ Ecorr of alloy, however, these particles do not sustain large
cathodic currents and may be too small to adversely impact corro-
sion kinetics. Pitting is not often associated with these intermetal-
lics.
Active particles with high self dissolution rates, e.g., MgZn2.
Ecorr Ͻ Ecorr of alloy, with the ability to undergo anodic dissolution
at high rates.
Active particles with low self dissolution rates. Not observed
here ͑although in a range of potentials, Mg2Al3 may qualify͒.
Active particles with a noble elemental component, e.g.,
Al2CuMg. Dealloying and incongruent dissolution may lead to po-
larity reversal. Can be selectively dissolved or lead to peripheral
pitting.
The survey presented here may be utilized either independently
or in conjunction with other localized measurement techniques to be
applied to the problem of alloy corrosion. This work focuses on one
specific aspect of localized corrosion in Al alloys, while in keeping
with the presentation of a survey of results, a deliberate effort has
been made not to discuss other factors such as local pH, transition
from metastable to stable pits, pit chemistry, etc.
3. F. Sato and R. C. Newman, Corrosion (Houston), 55, 3 ͑1999͒.
4. M. Zamin, Corrosion (Houston), 37, 627 ͑1981͒.
5. B. Mazurkiewicz and A. Piotrowski, Corros. Sci., 23, 697 ͑1983͒.
6. O. Seri, Corros. Sci., 36, 1789 ͑1994͒.
7. J. R. Scully, T. O. Knight, R. G. Buchheit, and D. E. Peebles, Corros. Sci., 35, 185
͑1993͒.
8. K. Nişancioğlu, J. Electrochem. Soc., 137, 69 ͑1990͒.
9. M. J. Pryor and J. C. Fister, J. Electrochem. Soc., 131, 1230 ͑1984͒.
10. A. Afseth, J. H. Nordlien, G. M. Scamans, and K. Nişancioğlu, Corros. Sci., 44,
2543 ͑2002͒.
11. J. L. Searles, P. I. Gouma, and R. G. Buchheit, Metall. Mater. Trans. A, 32A, 2859
͑2001͒.
12. G. O. Ilevbare and J. R. Scully, Corrosion (Houston), 57, 134 ͑2001͒.
13. G. O. Ilevbare and J. R. Scully, J. Electrochem. Soc., 148, 196 ͑2001͒.
14. C. M. Liao, J. M. Olive, M. Gao, and R. P. Wei, Corrosion (Houston), 54, 451
͑1998͒.
15. E. McCafferty, J. Electrochem. Soc., 150, 238 ͑2003͒.
16. M. Büchler, T. Watari, and W. H. Smyrl, Corros. Sci., 42, 1661 ͑2000͒.
17. G. O. Ilevbare, O. Schneider, R. G. Kelly, and J. R. Scully, J. Electrochem. Soc.,
151, 453 ͑2004͒.
18. O. Schneider, G. O. Ilevbare, R. G. Kelly, and J. R. Scully, J. Electrochem. Soc.,
151, 465 ͑2004͒.
19. T. J. R. Leclère and R. C. Newman, J. Electrochem. Soc., 149, 52 ͑2002͒.
20. R. G. Buchheit, L. P. Montes, M. A. Martinez, J. Michael, and P. F. Hlava, J.
Electrochem. Soc., 146, 4424 ͑1999͒.
21. L. F. Mondolfo, The Aluminum-Magnesium-Zinc Alloys, Research and
Development Center, Revere Copper and Brass Inc., Rome, NY ͑1967͒.
22. S. P. Ringer and K. Raviprasad, Mater. Forum, 24, 59 ͑2000͒.
23. Aluminum: Properties and Physical Metallurgy, J. E. Hatch, Editor, ASM, Metals
Park, OH ͑1984͒.
24. Q. Meng and G. S. Frankel, J. Electrochem. Soc., 151, 271 ͑2004͒.
25. J. A. Wert, Scr. Metall., 15, 445 ͑1981͒.
26. S. K. Maloney, K. Hono, I. J. Polmear, and S. P. Ringer, Scr. Mater., 41, 1031
͑1999͒.
27. M. Gao, C. R. Feng, and R. P. Wei, Metall. Mater. Trans. A, 29A, 1145 ͑1998͒.
28. Y. Yoon, PhD Thesis, The Ohio State University, Columbus, OH ͑2004͒.
29. N. Birbilis and R. G. Buchheit, Unpublished data.
30. R. G. Buchheit, M. A. Martinez, and L. P. Montes, J. Electrochem. Soc., 147, 119
͑2000͒.
31. V. O. Lavrenko, V. A. Shvets, S. O. Firstov, S. B. Prima, V. O. Kochubei, and V. M.
Adeev, Powder Metall. Met. Ceram., 42, 291 ͑2003͒.
Conclusions
32. R. G. Buchheit, R. K. Boger, M. C. Carroll, R. M. Leard, C. Paglia, and J. L.
Searles, JOM, 53, 29 ͑2001͒.
The information provided herein is a rich resource for clarifica-
tion pertaining to the electrochemical behavior of Al alloys. It has
been shown that the electrochemical microcell method was capable
of distinguishing between the electrochemical behavior of interme-
tallics, pure elements, and solid solutions common to Al alloys.
Overall we can begin to develop a classification system for in-
termetallics based on more than relative activity or nobility that
might be guessed from composition or measurement of corrosion
potential. This classification is based on intermetallic passivity or
activity, corrosion potential, corrosion current, and breakdown be-
havior.
33. M. Paljević, J. Less-Common Met., 175, 289 ͑1991͒.
34. R. M. Leard, M.Sc. Thesis, The Ohio State University, Columbus, OH ͑2001͒.
35. R. G. Buchheit, R. P. Grant, P. F. Hlava, B. McKenzie, and G. L. Zender, J.
Electrochem. Soc., 144, 2621 ͑1997͒.
36. C. M. Liao and R. P. Wei, Electrochim. Acta, 45, 881 ͑1999͒.
37. P. Leblanc and G. S. Frankel, J. Electrochem. Soc., 149, 239 ͑2002͒.
38. F. Andreatta, H. Terryn, and J. H. W. de Wit, Corros. Sci., 45, 1773 ͑2003͒.
39. F. Andreatta, H. Terryn, and J. H. W. de Wit, Electrochim. Acta, In press ͑2004͒.
40. S.-M. Moon, M. Sakairi, and H. Takahashi, J. Electrochem. Soc., 151, 399 ͑2004͒.
41. J. C. Seegmiller and D. A. Buttry, J. Electrochem. Soc., 150, 413 ͑2003͒.
42. F. Andreatta, M. M. Lohrengal, H. Terryn, and J. H. W. de Wit, Electrochim. Acta,
48, 3239 ͑2003͒.
Acknowledgments
The assistance of Douglas Pohlmann and Daniel Huber with
sample preparation, Nikki Padgett with the microcell, and Cameron
Begg with microanalysis, are gratefully acknowledged.
43. T. Suter and H. Böhni, Electrochim. Acta, 43, 2843 ͑1998͒.
44. T. Suter and H. Böhni, Electrochim. Acta, 47, 191 ͑2001͒.
45. M. M. Lohrengal, Corrosion Engineering Science and Technology, 39, 53 ͑2004͒.
46. T. Suter and R. C. Alkire, J. Electrochem. Soc., 148, 36 ͑2001͒.
47. E. McCafferty, Corros. Sci., 37, 481 ͑1995͒.
The Ohio State University assisted in meeting the publication costs of
this article.
48. Corrosion of Aluminum and Aluminum Alloys, J. R. Davis, Editor ASM, Materials
Park, OH ͑1999͒.
References
1. I. J. Polmear, Light Alloys, 3rd ed., Arnold, London ͑1995͒.
49. Z. Szklarska-Smialowska, Corros. Sci., 41, 1743 ͑1999͒.
Downloaded on 2014-12-08 to IP 130.194.20.173 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).