Materials Science s a nd Applications, 2011, 2, 1542-1555
doi:10.4236/msa.2011.211207 Published Online November 2011 (http://www.SciRP.org/journal/msa)
Copyright © 2011 SciRes. MSA
Corrosion Behavior of Heat Treated
Nickel-Aluminum Bronze Alloy in Artificial
Seawater
Ashkan Vakilipour Takaloo1, Mohammad Reza Daroonparvar2, Mehdi Mazar Atabaki1,3,
Kamran Mokhtar1
1Department of Materials Engineering, Faculty of Mechanical Engineering, Universiti Teknologi Malaysia, Johor, Malaysia; 2De-
partment of Materials Engineering, Faculty of Mechanical Engineering, Roudehen Branch, Islamic Azad University, Roudehen, Te-
hran, Iran; 3Institute for Materials Research, School of Process, Environmental and Materials Engineering, Faculty of Engineering,
University of Leeds, Leeds, UK.
Email: pmmmaa@leeds.ac.uk, m.mazaratabaki@yahoo.co.uk
Received June 28th, 2011; revised August 11th, 2011; accepted September 14th, 2011.
ABSTRACT
The effect of microstructure of nickel-aluminum bronze alloy (NAB) on the corrosion behavior in artificial seawater is
studied using linear polarization, impedance and electrochemical noise tests. The alloy was heat treated in different
heating cycles including quenching, normalizing and annealing. Microstructure of the specimens was characterized
before and after heat treatment by optical microscopy and scanning electron microscopy. Results showed that the value
of pearlite phase in the normalized alloy is much more than other specimens, leading to higher corrosion resistance.
Polarization test showed that starting point of passivation in the polarization of the normalized alloy is lower than other
specimens. The dissolution of Mn and Fe rich phases increased the Mn and Fe contents in solid solution, and this en-
hanced the passivation power of the surface of the alloy. The effect of the alloying elements was seen by a lower corro-
sion potential and an inflexion at around 280 mV (SCE) in the polarization curve, indicating the preferential dissolution
of some elements beyond that potential. The polarization curve showed that the anodic polarization behavior of the al-
loy in the solution was essentially controlled by the intermetallic phases, mainly containing Cu. Two types of corrosion,
pitting and selective corrosion, were observed in the specimens after being exposed to artificial seawater.
Keywords: Nickel-Aluminum Bronze Alloy, Heat Treatment, Corrosion, Microstructure
1. Introduction
Nickel-aluminum bronze known as NAB is a series of
copper-based alloy with additions of 9% - 12% Al and
6% Ni and Fe. High corrosion resistance of this alloy has
made it one of the most practical alloys in marine appli-
cations e.g. ship propellers [1,2]. The microstructure of
the alloy consists of a Cu-rich solid solution known as
α-phase and β′-phase or martensitic β-phase, surrounded
by lamellar eutectoid phase and a series of intermatallic k
phases [3 and 4]. Among the inermetallic compounds, KI
phase is rosette shape which is rich in Fe, KII phase is
smaller than kI phase and form a dendritic rosette shape
which distributed at the α/β boundaries, KIII phase is la-
mellar shape and it forms at the boundary of KI phase and
is rich in Ni and KIV phase is a fine Fe rich precipitations
that forms in α phase [5]. Recently, a vast range of inves-
tigation have been carried out to study the corrosion be-
havior of the cast nickel-aluminum alloy [6,7] and it has
been found that optimum corrosion resistant of the alloy
in seawater can be obtained by controlling the mictro-
structure [8]. Corrosion of the bronze alloy in seawater
have been investigated and the tendency of the cast alloy
to corrode was attributed to the formation of β′ phase
which plays an anodic role compare to α matrix [9]. The
complexity of the alloy with several intermetallic phases
lead to a significant change in the development of the
microstructures, which can result in extensive corrosion
resistance in seawater. Results of salt spray test on corro-
sion behavior of the alloy indicated that corrosion resis-
tance has been improved in the heat treatment sequences
like aging, quenching, normalizing and annealing. Char-
acterizing the microstructure showed that quenching
transforms all β phase into β′ phase and aging results in
Corrosion Behavior of Heat Treated Nickel-Aluminum Bronze Alloy in Artificial Seawater1543
precipitation of fine k phases. On the other hand, anneal-
ing has lead to the transformation of β′ martensite into α
and k phases [10]. Microstructure of the bronze alloy
suffered from cavitation corrosion in seawater showed
that the alloy was corroded by selective corrosion at the
interface between α phase and intermetallic K precipi-
tates, however, K precipitate-free zones were not consid-
erably corroded. Crevice corrosion of the copper-based
alloy was often associated with the formation of a Cu-ion
concentration cell and an accelerated attack at the ex-
posed area adjacent to the crevice [11]. Meigh et al. [12]
declared crevice corrosion occurs in the nickel-aluminum
bronze when it is not cathodically protected. It has been
reported that the rate of crevice corrosion of the alloy in
seawater is about 0.7 - 1.0 mm y-1 [13,14]. The corro-
sion resistance of the alloy is owned to a protective layer
containing both Al and Cu oxides, with 900 to 1000 nm
thickness [15,16]. The Al-rich oxide layer usually forms
in the vicinity of the alloy and the Cu-rich forms in the
external regions. Oxides of Ni and Fe, and a little Cu
salts and Cu hydroxychlorides (Cu2(OH)3Cl and Cu
(OH)Cl) exist on the surface of the alloy after exposing
for long time to seawater. These oxide layers provide
corrosion protection by mutually lessening the anodic
dissolution reaction, stopping the ionic transport across
the oxide layers, and a decrease in the rate of the ca-
thodic reaction on the oxide layers. On the contrary to the
many studies of the corrosion in seawater for the alloy,
there has been small number of investigations to study
the effect of heat treatment processes on the electro-
chemical kinetics at the surface of the alloy when it is
exposed to seawater environment [17]. For example,
Schüssler and Exner [11] conducted a research on the
dissolution of a cast nickel-aluminum bronze alloy in
seawater with considering a fixed velocity of electrolyte
without a considerable contribution on the effect of mi-
crostructure on the corrosion. Kear et al. showed that
both cathodic [16] and anodic [17] polarization can occur
during the corrosion of the alloy in an aqueous solution.
The anodic dissolution of the alloy showed significant
dissolution of Cu with forming dichlorocuprous anion
and the cathodic reaction showed four electron reduction
of dissolved oxygen on the surface. However, at a mixed
potential, the anodic kinetics are controlled by mass
transfer and charge transport. As a result, the rate and
mechanism of the corrosion are very susceptible to prior
heat treatment processes. The protective nature of the
alloy was attributed to the formation of a Cu2O layer due
to the history of its heat treatment. The reduction of the
corrosion rate during the layer formation process was
explained by means of the concurrent decrease both of
the anodic reaction inside the growing Cu2O layer, and
the cathodic reaction on its surface. Aluminum in pas-
sivation was incorporated in the Cu2O lattice, reducing
the rate of oxygen reduction on the Cu2O surface when
the alloy exposed to a corrosive environment [17].
From another stand point, electronic tools like linear
polarization have some drawbacks for the measurement
of electrochemical properties. One of the main problems
is the invasive character of the devices to the electro-
chemical systems of the metallic electrodes in aqueous
solutions [18]. In this regard, the corrosion current den-
sity of metallic electrodes, corresponding to the open-
circuit potential of the electrodes in the aqueous solution
can be measured with zero resistance ammeter and linear
polarization methods [19]. In zero resistance ammeter the
corrosion current can be estimated with measuring cur-
rent density of two similar metallic electrodes at the
open-circuit potential of the electrodes in seawater with-
out any external voltage on the electrodes. This process
owned to the existence of electronic noise and loss of the
electrical resistance across the interface between two
electrodes and seawater. However, in the linear polariza-
tion method, the corrosion current density of electrode at
the open-circuit potential of the electrode in artificial
seawater can be measured with an external voltage. In
this process, polarization resistance expands across the
interface between the solid electrode and seawater. As
one part of the present study concentrates on the results
of electrochemical noise technique, the fluctuations of
the current and potential were generated at the same time
during the corrosion process. The most effective part of
this test was the elimination of disturbing signals and the
ability to avoid the artificial disturbances during the
measurement. It was shown that the sensitivity of elec-
trochemical noise measurement is considerably higher
than other conventional methods in the case of localized
corrosion process [20]. In this method, at different times
the intensity of corrosion can be monitored and the noise
resistance, the ratio of the standard deviations of the po-
tential and current noise can be easily measured. The
mechanism of the process is a competition course be-
tween breakdown and repassivation of the passive film
with developing a pitting and an active corrosion process.
In this state, forming unstable pits depends on the time
and repassivation is the consequence of the shape and life
of the unstable pits. For estimating power spectral den-
sity by the maximum entropy method, it is essential to
choose the maximum order of the autoregressive acci-
dental variable. If the order is small, the spectrum can be
approximately flat and if the spectrum emerges noisier it
shows the order is large [21]. In the statistical calculation
of the noises, the highest noises called shot-noise were
analyzed and the individual occurrences in the corrosion
process were considered.
In the present study, the influence of different pre-heat
Copyright © 2011 SciRes. MSA
Corrosion Behavior of Heat Treated Nickel-Aluminum Bronze Alloy in Artificial Seawater
Copyright © 2011 SciRes. MSA
1544
treatments on the passivation behavior of nickel-alumi-
num bronze alloy is studied using linear polarization,
impedance and electrochemical noise tests. The effect of
heat treatment processes during corrosion of the alloy
was evaluated by optical microscopic, scanning electron
microscopy and X-ray diffraction analysis. It was found
that normalizing of the alloy has a considerable effect in
preventing the corrosion in artificial seawater.
2. Experimental Details
The nickel-aluminum bronze alloy (C95520) was cast
according to ASTM B 148 standard and then cut into
small pieces of dimensions 6 mm × 6 mm × 3.5 mm. The
composition of the alloy was determined by X-ray fluo-
rescence spectrometry (XRF) and it is shown in Table 1.
Tensile strength and hardness of the substrate esti-
mated experimentally shown the alloy has tensile
strength of about 630 MPa and hardness of 162 HV. The
phase composition has been checked at high exposure
times by X-ray diffractometry (XRD) in the Bragg-
Brentano geometry, using Cu = Kα radiation (k =
0.15404 nm). The specimens were mechanically cleaned,
degreased, immersed in an aqueous solution of 20%
HNO3 and then dried and maintained for 24 h in a desic-
cator in excess of silica gel and in the last part of the
cleaning process the alloy weighed. For preparing the
aqueous solution, artificial seawater was made according
to ASTM D 1141-98 standard. The combination of the
compounds in the artificial seawater is given in Table 2.
The artificial seawater with pH of 8.26, dissolved
oxygen concentration of around 7.0 ppm and conductiv-
ity of 59.8 mS m–1 was held at ambient temperature of
27˚C during the test. Before accomplishing any corrosion
test, the specimens were heat treated in different cycles.
The applied heat treatment procedures for altering the
microstructure of the alloy are given in Table 3.
The specimens were mechanically polished with 0.2
μm α-alumina slurry and etched with the solution of 5 g
FeC13 + 25 mL HC1 + 50 mL H20. X-ray diffraction test
was performed to realize the intermetallic compounds in
the heat treated specimens before and after corrosion
tests. To perform the polarization test, the specimens
were grounded, polished and degreased with acetone. To
prevent other areas from exposing to seawater an area
with a dimension of 1 mm × 1 mm was chosen and other
areas were protected with a polymeric adhesive. The ad-
hesive prepared by combining cure polyamide and rein-
forcement shielded with BissWax in order to minimize
the localized corrosion. General corrosion behavior e.g.
oscillation in corrosion potential of the alloy was esti-
mated using linear polarization with EG & G, model
237A corrosion system. The data recorded from this test
was analyzed by MATLAB V7.0.0 software. The poten-
tial scanning rate was 0.5 mVs–1 and when the potential
reached to 800 mV, the scan was discontinued. In this
test, Ag/AgCl and Pt reference electrodes were used in a
potential between –700 and 800 mV setting in maximum
value for 0.6 second. The platinum counter electrode
enabled controlling the potentiostatic during the test. The
recorded values of corrosion rate correspond to the aver-
age value from the specimens per exposure time. The
passivation resistance layer was estimated with imped-
ance test according to ASTM G 106-89 standard. In the
test, potansiostat EG & G, model 273A device was em-
ployed and M398 V.1.30 software was used for re-
cording impedance data and drawing Bode curve during
the test. To record the data from Nyquist test, Z
view-V3.0c software was employed. After linear polari-
zation and impedance test, scanning electron microscope
(SEM) was used for characterizing the microstructural
changes. In electrochemical noise measurements, 22
small containers with a capacity of 400 cc were prepared.
The specimens were kept in 1cm distance from each
other in the container with protecting the environment
from any vibration or pulsation. The anode to cathode
area ratio was 1:1. The specimens were connected via
electrical leads, with a gap of 1 cm and were mounted in
a planar orientation on two non-metallic boards. The
boards were immersed in artificial seawater. The current
and potential were measured using a zero resistance am-
meter, model SNAWA. Formation and destruction of
passive film on the surface of specimens, initiated by
pitting corrosion, was recorded by InstLinkPC software.
The reference electrode was located in a free space be-
tween two of the specimens with the same applied
pre-heat treatment process. Specimens held in artificial
seawater for periods of 11, 20, 40, 50 and 75 days were
tested in electrochemical noise examination (ZRA) for
1400, 2500, 4000, 4500 and 6000 second. Figure 1
shows how specimens were located in the container for
corrosion tests.
Table 1. Chemical composition of nickel-aluminum bronze alloy (wt%) after casting.
Element Composition (wt%)
Material
Cu Al Fe Ni Si Pb Zn Sn Mn
Nickel-aluminum bronze alloy (NAB) Bal. 11.54 4.47 4.60 0.030 0.017 0.002 0.005 0.001
Corrosion Behavior of Heat Treated Nickel-Aluminum Bronze Alloy in Artificial Seawater1545
Table 2. Chemical composition of artificial seawater based
on ASTM D 1141-98 standard.
Component Quality Unit
NaHCO3 0.201 g/L
CaCl2 1.158 g/L
MgCl2.6H2O 11.112 g/L
SrCl2.6(H2O) 0.042 g/L
NaCl 25.534 g/L
Na2SO4 4.094 g/L
KCl 0.695 g/L
KBr 0.201 g/L
H3BO3 0.201 g/L
NaF 0.201 g/L
Table 3. Applied heat treatment cycles to nickel-aluminum
bronze alloy.
Hear treatment Group
name
Time
(min)
Specimen
code
Temperature
(°C)
1 475
9 675
Quenching in
water, 27°C A 15
17 825
3 475
11 675
19 825
Quenching in
water, 27°C B 30
25 900
5 475
13 675
Quenching in
water, 27°C C 45
21 825
7 475
15 675
Quenching in
water, 27°C D 60
23 825
2 475
10 675 Normalizing in air E 15
18 825
4 475
12 675
20 825
Normalizing in air F 30
26 900
6 475
14 675 Normalizing in air G 45
22 825
8 475
16 675
Normalizing in air H 60
24 825
Figure 1. Schematic of the location of the alloy and elec-
trodes in the container filled with artificial seawater.
3. Results and Discussion
The microstructure is characterized by metallographic
techniques and is found to consist of α-phase, retained
β-phase and numerous K-phases. The microstructures of
the specimens before and after heat treatment are shown
in Figures 2 and 3, respectively. In the cast alloy, β
phase transforms to α phase together with some lamellar
KIII at grain boundaries and globular star shape KII pre-
cipitates. KIII is Ni based (usually AlNi) and KII and KIV
are a combination of Fe and Al was always emerging as
Fe3Al compound in the microstructure (see Figure 2).
Normalizing at a lower temperature causes formation
of fine Fe rich KIV within the grains. However, if the
cooling was not slow enough, some of the high tempera-
ture β was retained as a martensitic structure. The mart-
ensite was then transformed into a very fine mixture of α
and K phases, with NiAl precipitates, referred to as tem-
pered martensite (see Figure 3).
XRD analysis was performed in one of the specimens
normalized at 900˚C showing Cu-rich α phase was
formed along with KIII or KIV in the matrix (Figure 4). It
was observed that with increasing the normalizing tem-
perature β phase increased dramatically and eutectoid
transformation products together with K phases were
decreased. This may in turn affect the reaction products
with the corrosive environment. Figure 5 shows the mi-
crostructure of the alloy after normalizing at 900˚C
which KIII and KIV phases were formed in the matrix as
well as presenting retained β phase.
However, the quenched alloy contained Widmanstat-
ten and α phase in a matrix of martensite. When the
specimens were cooled in air more Widmanstatten and α
have formed in the matrix. Additionally, the specimen
held in longer time at the heat treatment temperature and
quenched in water contained martensite and bainite plus
fine eutectoid surrounding the Widmanstatten and α. For
Copyright © 2011 SciRes. MSA
Corrosion Behavior of Heat Treated Nickel-Aluminum Bronze Alloy in Artificial Seawater
1546
(a)
(b)
SEM MAG:3.00 KX Det:SE Detector
SEM HV:15.00 KV WD:23.6120 mm
Vac: HiVac
VEGA/TESCAN
10 μm
Figure 2. (a) Microstructure of the alloy before heat treatment and corrosion tests and (b) star shape intermetallic at the cen-
tre of the non-treated specimen.
50 μm
50 μm
(a) (b)
50 μm
50 μm
(c) (d)
Figure 3. Microstructure of the alloy after normalizing at (a) 475; (b) 675; (c) 825 and (d) 900˚C for 30 min.
Copyright © 2011 SciRes. MSA
Corrosion Behavior of Heat Treated Nickel-Aluminum Bronze Alloy in Artificial Seawater
Copyright © 2011 SciRes. MSA
1547
Figure 4. X-ray diffraction from the specimen normalized
at 900˚C.
SEM MAG:2.00 KX Det:SE Detector
SEM HV:15.00 KV WD:23.4690 mm
Vac: HiVac
VEGA//TE SCAN
20 μm
Figure 5. Microstructure of the alloy normalized at 900˚C.
example in the specimen of group A in Table 3, temper-
ing transformed martensite and bainite to a fine eutectoid.
As it is seen in Figure 5, when the alloy was cooled in
air, Widmanstatten and α phase, eutectoid and traces of
bainite and martensite were obviously appeared in the
microstructure. However, when the specimens were
normalized at 900˚C, some of β phase transformed into β′
phase, and the other transformed into lamellar α and K
intermetallic phases distributing along the boundaries of
the alloy. Although martensitic transformation happened
during quenching, the microstructure of the alloy still
consisted of α, β′ and K phases. It was seen that the vol-
ume fraction of β′ phase increased in the quenched
specimens due to the dissolution of both α and K phases.
3.1. Linear Polarization Test
Figure 6 shows the result of potentiodynamic polariza-
tion for the alloy as a function of heat treatment cycle,
measured using linear polarization method. The results
show a significant similarity in the polarization curves of
the specimens heat treated at different heating cycles. It
is noted that for the specimen normalized at 675°C for 45
min and then cooled in air (code 14) the polarization re-
sistance was higher than other specimens and its related
curve in the polarization diagram shifted to the left side.
The results shown in Figure 6 may reflect that the alloy
exhibits a higher protective phases in the matrix in the
corrosive environment. Although the effect of micro-
structural difference between intermetallic compounds,
primary phases and grain boundary on the corrosion be-
havior of the metal is not yet well established, the greater
polarization resistance of the specimen with the men-
tioned condition may indicate that some phases are more
resistant to corrosion.
The faster cathodic reaction, resulting from a micro-
structure with less corrosive phases, was established to
associate with the reaction of Fe and Cu atoms with other
elements in imperfection sites introduced by the heat
treatment cycles. From another point of view, in the alloy
subjected to air cooling from high temperatures, a higher
dislocation density is anticipated, causing a raise in ca-
thodic reaction rate [22]. Similarly, for the coarsened
grained alloy heat treated in lower temperature, the
coarsening of Widmanstatten phase produces lower dis-
location density, leading to a reduction in cathodic reac-
tion rate and a superior polarization resistance. Further-
more, the presence of Fe showed improvements in corro-
sion behavior of the alloy. Williams and Komp [23] have
reported that an increase in Cu content affected the ca-
thodic polarization behavior in such a way as to reduce
the rate of the cathodic reaction in the corrosion process,
thereby decreasing corrosion rate. The polarization resis-
tances of the heat treated specimens, however, were
higher than that of the non-heat treated alloy. It is obvi-
ous that higher Al and Cu contents in the alloy provided
by impoverishing some phases offer the beneficial effect
on the corrosion resistance as the heat treated alloy (code
14) exhibits greater protective layers on the surface.
As it was seen, an especial heat treatment can affect
the results of linear polarization in a way that some of
them showed high electronic and ionic resistance during
the immersion, indicating the products of the corrosion
system are more stable and protective. Analysis of the
linear polarization curves revealed active-passive behav-
ior for the alloy. After the passive range, rapid increases
of current density occurred and passive layer destruction
proceeds and following transition into pitting corrosion
region is seen for the alloy heat treated with the quench-
ing process. Before this transition, some ephemeral pits
formed but were rapidly re-passivated. The main corro-
sion in this case can be pitting. In contrast, for the sub-
strate subjected to normalizing, icorr increased smoothly
with increasing potential and the usual active-passive
transition was not appeared. This can be attributed to
Corrosion Behavior of Heat Treated Nickel-Aluminum Bronze Alloy in Artificial Seawater
1548
Figure 6. Linear polarization test results for the alloy heat treated in different conditions.
selective corrosion. It also conceived that commence of
passivation was from lower potentials indicating superior
specimens for preventing corrosion in artificial seawater
solution. In addition, the normalized specimens exhibited
more extended passive behavior in comparison with the
quenched specimens. It was also found that the K phases
were not attacked, indicating that they were cathodic
with respect to the Cu-rich α phase. The continued dis-
solution of α phase is expected to result in the eventual
loss of the K intermetallic compounds. However, it was
reported that the corrosion of the alloy in artificial sea-
water is basically galvanic in nature [24,25]. The Cu-rich
phase was selectively attacked at the interfaces with the
intermetallic phases in artificial seawater. Cavitations
resulted in a rough surface, containing large size cavities,
ductile tearing and grain boundary attack. Microcracks of
about 12 mm in length were observed in Cu rich phase
adjacent to intermetallic phases. It is believed that selec-
tive phase corrosion and cavitation stresses were the
main cause of cracking [26]. However, a big doubt about
the mechanism of corrosion in the specimens has spurred
this investigation to be extended. Hence, extracted results
of potentiodynamic curves by Tafel extrapolation are
reported in Figure 7. The main results derived from the
performed potentiodynamic tests are repeatedly higher
corrosion resistance of the normalized specimens com-
pared to the cast and quenched specimens. Noticeably,
the superior specimen made up a balance between low
current density, low potential and high passive range.
When martensite is considered, a high energy level
seems to be associated with its resulting microstructure
which can significantly affect its corresponding electro-
chemical behavior. Due to this, it is expected a higher
susceptibility to corrosion than those other examined
heat-treated specimens. Nonetheless, both martensite
structure and volume fraction of dendritic Cu-rich phase
were strongly affected by the applied cooling rate intrin-
sically in each process.
The effect of heat treating cycles on the polarization
characteristics of the alloy depends on the composition of
the alloy. With the presence of Cu and Al in the micro-
structure the polarization has been significantly modified
via dissolving the Cu and Al and increasing the Cu con-
tent in solid solution. In the alloy, the dissolution of the
Ni and Cu rich phases increased the Al contents in solid
solution, but did not significantly enhances the passiva-
tion power of the surface of the alloy. The similarity be-
tween the polarization curve for the reference specimen
and those heat treated seems to support this point. To
substantiate the point, the polarization curve for a refer-
ence specimen was acquired. By comparison, it can be
seen that the curve for the reference specimen and the
heat treated specimens were very similar. This point is
clearer in the peak current density and the corresponding
potential, and also the passive current density, indicating
that the anodic polarization behavior of the alloy in the
aqueous solution was fundamentally controlled by Cu-
and Al-containing phases. Nonetheless, the effect of the
Copyright © 2011 SciRes. MSA
Corrosion Behavior of Heat Treated Nickel-Aluminum Bronze Alloy in Artificial Seawater1549
Figure 7. Various corrosion current rate after polarization test.
heat treatment cycles was obvious by a lower corrosion
potential and an inflexion at around 600 mV (SCE) in the
polarization curve, indicating the preferential dissolution
of some elements beyond that potential. It should be
mentioned that the differences in Tafel results are attrib-
uted to differences in the surface film composition which,
in turn, arise from the compositional differences between
the alloys heat treated in different cycles.
The importance of the effect of the microstructures on
the corrosion behavior of the alloy was a significant in-
dication of the influence of heat treatment cycles on the
surface morphology evolution. For instance, the signs of
corrosion were observed at a very early stage with
roughening of the KI phase. Local attack, however, at the
α/KI boundaries began to take place after six days (see
Figure 8).
With increasing the immersion time, corrosion prod-
ucts began to appear in α and β matrix, but the corrosion
attack was much more rigorous at the active K phases
and α/K boundaries. This attack was signaled delineation
of grain boundaries at the first stages of corrosion.
Moreover, some corrosion products began to appear at
the grain boundaries of the specimens normalized at
675˚C. Normalizing process reduced the chance of
forming local galvanic cells which might exist between
different phases, and improved the corrosion resistance
of the alloy. The calculated corrosion potential Ecorr for
all specimens indicated a close range between the results
leading to an average estimation of about -305/66 mv.
For easing the corrosion analysis, some specimens (ref.,
3,5,6,12,14,18,20,25 and 26) with highest and lowest icorr
were chosen for electrochemical impedance test from the
groups shown in Table 3. Each experiment was carried
out two times in order to produce more reliable results.
The solution resistance (RS) and the polarization resis-
tance or charge transfer resistance (Rp) were derived us-
ing the equivalent circuit model as shown in Figure 9.
As it is seen in the model, R1 is equivalent to Rs. Fig-
ure 10 shows the Nyquist plots obtained from AC im-
pedance (EIS) measurement for the specimens immersed
SEM MAG:2.00 KX Det:SE Detector
SEM HV:30.00 KV WD:22.5450 mm
Vac: HiVac
VEGA//TESCA
N
20 μm
Figure 8. SEM micrographs showing evolution of damage
morphology in the specimen normalized at 675˚C and im-
mersed in artificial seawater for six days.
Copyright © 2011 SciRes. MSA
Corrosion Behavior of Heat Treated Nickel-Aluminum Bronze Alloy in Artificial Seawater
1550
in artificial seawater. For plotting the diagrams the cor-
rosion rate (CR) and corrosion current density (iCorr) were
calculated from Equations (1) and (2) in a code written in
the software.
2
1mA
46 cm
Corr
p
iR

(1)
1300 mpy
Corr
R
Mi
CnD

(2)
where, D, M and n are density (g/cm3), molecular weight
(g/mol) and valence of the alloy, respectively. As can be
seen from Figure 10, different behavior can be observed
considering all parameters involved in the test. The
amount of the polarization resistance taken from the EIS
measurement in the alloy depends on the history of heat
treatment of the specimens. For instance, RP value of the
normalized specimens at 675˚C enhanced in comparison
with the quenched specimens.
By comparing the results obtained from different
specimens, it can be seen that the Nyquist plots have
small polarization resistances at the open circuit poten-
tials. This means that charge transfer process controls the
corrosion behavior of the specimens in artificial seawater.
The Nyquist plot also clearly shows that the quenching
process caused displacing in the left side of the Nyquist
plots. It means that the quenching treatment led to a de-
crease in the impedance of the specimens. However, it is
interesting to note that, normalizing the specimens
changes the diameter of the semicircles in the Nyquist
plots indicating an increase in the value of polarization
resistance. Based on the results from the EIS measure-
ment, it can be stated that the normalized alloy (specimen
14) showed the highest corrosion resistance, Rp, com-
pared to other specimens. It should also declare that in-
significant changes occurred in the Nyquist plots of some
substrates. This might be attributed to the effect of mi-
crostructure of the heat treated alloys for absorbing arti-
ficial seawater due to dissociation into anions and cations
and their involvement to the electrochemical process
with increasing the corrosion rate [27].
Among the specimens the corrosion rate in the nor-
malized alloy decreased to around 19%. This result can
probably be attributed to the supply of oxygen for the Al
passivation by artificial seawater and formation of a com-
Figure 9. Equivalent electrical circuit model used to derive
RS and Rp.
pact thin layer of Al oxide on its surface and finally
reaching to Warburg impedance. However, in the nor-
malizing process if the temperature reached to 900°C the
alloy lost its protection against the corrosive environment.
This can be associated with the dissolution of the oxide
layer and Cu-rich phase as anodic walls during the corro-
sion process. The Warburg impedance was created with
diffusion process. As the impedance depends on the fre-
quency of the potential perturbation, at high frequencies
the Warburg impedance was small since diffusing reac-
tants did not have to move very far. At low frequencies
the reactants had to diffuse farther, increasing the War-
burg-impedance. On the Nyquist plot the Warburg im-
pedance appeared as a diagonal line with a slope of 45°
and on the Bode plot, the Warburg impedance exhibited
a phase shift of 45˚. At the higher normalizing tempera-
ture passivation layer resistance had faced with diffu-
sional phenomenon. It means that another resistance cir-
cuit, a new one (Zw), added to the electrical circuit. This
can be shown in the electrical circuit with Rp. It is prob-
able the initial time taken to achieve the stabilized poten-
tial is associated with the dissolution and conversion of
the thermal oxide films. The change from more positive
potentials exhibited by the oxides to more negative po-
tentials is due to the changes in the surface films. This
process relies on the rate of mass transport for the chlo-
ride ion and the cuprous dichloride complex, both of
which are engaged in the dissolution of the cuprous oxide.
The passivation of the alloy was based on the oxidation
of aluminum. It is believed that the aluminum has an
enormous affinity for oxygen than other solutes like Cu
and always Al2O3 is approximately ten times more stable
than Cu2O [5]. However, a single equivalent circuit can
be applied where Rs is the solution resistance which in
Bode diagram is expressed in a high frequency limit (F>1
Hz), R1 is the passive oxide film resistance and ZCPE is
the constant-phase element for the oxide film. On the other
hand, as other alternative a complex equivalent circuit
consisted with Rs also corresponding to the solution re-
sistance, R1 and R2 are the resistances of the porous and
barrier layers, respectively [28-31]. For the substrate,
thermal oxidation was based on a rapid initial production
of Cu2O at the alloy/oxide interface due to the depletion
of Cu. Alumina consequently formed as a protective ox-
ide which was extremely resistant to the passage of cu-
prous cations which can no longer enter in the layer of
cuprous oxide. Providing higher aluminium content by
an appropriate thermal cycle can lead to the greater cor-
rosion resistance due to the protective alumina film.
3.2. Electrochemical Noise Measurements
Electrochemical noise analysis shown in Figure 11 con-
irmed that the type of corrosion in the specimens nor- f
Copyright © 2011 SciRes. MSA
Corrosion Behavior of Heat Treated Nickel-Aluminum Bronze Alloy in Artificial Seawater
Copyright © 2011 SciRes. MSA
1551
(a)
3 N
5N
6 N
12 N
14 N
15 N
18 N
20 N
25 N
26 N
AEFN
(b)
(c)
Figure 10. (a) Bode plots, (b) Nyquist plots of the specimens tested in artificial seawater based on the heat treatment histories,
obtained from electrochemical impedance spectroscopy test and (c) three dimensional plot relating all parameters involved in
the test.
Imaginary component of impedance
I(Zim. Cm2)
Imaginary component of impedance
2) I(Zim. Cm
Imaginary component of impedance
I(Z . Cm2)
im
Corrosion Behavior of Heat Treated Nickel-Aluminum Bronze Alloy in Artificial Seawater
1552
malized at 675˚C was selective and pitting corrosions.
For the normalized specimens, it was observed that after
20 days immersing in artificial seawater there was not
any considerable change in current and potential noises.
After reaching to 40 days, a slight drop in potential to
negative volumes was observed. This was attributed to
the nucleation of semi-stable pits on the passive surface
in selective corrosion mechanism. Transmuted noises
from high frequencies with low interval in current noise
indicated dealominification was the selective phase at the
primary stage of the corrosion process. After 50 days,
alterations in noise were dramatically increased and the
passive surface lost its ability to provide essential current
for the selective corrosion. Consequently, after 75 days
the current was extensively decreased. At this stage the
possibility of forming unstable pits were minimized.
However, a sharp drop in current was seen indicating a
dissolution or consumption of Ni, Al and Fe in the pas-
sive film. This phenomenon formed bubble shape corro-
sion products from Cu rich passive surface and increased
Figure 11. ZRA curves for the normalized ANB alloy after (a) 11 days; (b) 20 days; (c) 40 days; (d) 50 days and (d) 75 days.
Copyright © 2011 SciRes. MSA
Corrosion Behavior of Heat Treated Nickel-Aluminum Bronze Alloy in Artificial Seawater1553
anodic areas. A comparison between different specimens
shows dissimilar response in terms of potential and cur-
rent via immersing days in ZRA test (Figure 12).
As it is seen in Figure 12, the highest value of poten-
tial (–123/4631 mV) is for specimen 14, the one normal-
ized at 675˚C. The reference specimen also showed a
current value of 6/148 μA, while specimen 14 shows the
lowest current value of about 3/1348 μA after 75 days
immersing in artificial seawater. The results indicated
that specimen 14 was heat treated in an optimum condi-
tion among other heat treated nickel-aluminum bronze
specimens. In this view, lower corrosion Fe-base prod-
ucts were appeared in the specimen 14. The anodic reac-
tion in the retained β matrix was lower than other phases
in this specimen leaving the phase uncorroded. At 675˚C
of normalizing temperature, the temperature was not
enough to increase the current corrosion while in other
specimens, especially those normalized at 900˚C, the
current corrosion increased. SEM image of the surface of
the reference specimen after exposing to artificial sea-
water for 75 days is shown in Figure 13. In the alloy heat
treated with quenching and normalizing processes some
(a)
(b)
Figure 12. Comparison between current and potential of
ZRA test for the specimens normalized and quenched at
different temperatures for 45 min.
SEM MAG:5.00 KX Det:SE Detector
SEM HV:15.00 KV WD:22.9060 mm
Vac: HiVac
VEGA// T ESC A
N
5 μm
Figure 13. SEM images from the surface of reference spe-
cimen after exposing to artificial seawater for 75 days.
of the elements were dissolved in the passive layer mak-
ing some parts more anodic than other parts. This pre-
pares some local areas for aggressive pitting corrosion.
4. Conclusions
The influence of pre-heat treatment on the corrosion be-
havior of nickel-aluminum bronze was investigated. The
following results obtained from the present study can be
drawn:
1) The normalized specimens exhibited significantly
lower fractions of pits and better corrosion resistance to
artificial seawater in comparison with quenched speci-
mens. Superior specimen showed better corrosion resis-
tance in both short and long term corrosion tests, even
better than reference specimen.
2) The corrosion rate of the alloy was dependent on
many parameters including the nature of the surface
films, the heat treatment process, the alloy composition
and the time of immersion in artificial seawater.
Polarization plots showed that the Warburg impedance
occurred in the passive layer of the specimen normalized
at 675˚C. The corrosion resistance of the alloy benefited
from the microstructure resulted from pre-heat treatment
process.
REFERENCES
[1] R. C. Barik, J. A. Wharton, R. J. K. Wood, K. S. Tan and
K. R. Stokes, “Erosion and Erosion-Corrosion Perform-
ance of Cast and Thermally Sprayed Nickel-Aluminium
Bronze,” Wear, Vol. 259, 2005, pp. 230-242.
do i:1 0. 10 16 / j. wear. 2 0 05 .02 .033
[2] A. Al-Hashem and W. Riad, “The Role of Microstructure
of Nickel-Aluminium-Bronze Alloy on its Cavitation
Corrosion Behavior in Natural Seawater,” Materials
Copyright © 2011 SciRes. MSA
Corrosion Behavior of Heat Treated Nickel-Aluminum Bronze Alloy in Artificial Seawater
1554
Characterization, Vol. 48, 2002, pp. 37-41.
doi:10.1016/S1044-5803(02)00196-1
[3] F. Hasan, J. Iqbal and N. Ridley, “Microstructure of As-
cast Aluminum Bronze Containing Iron,” Materials Sci-
ence Technology, Vol. 1, 1985, p. 312.
[4] F. Hasan, A. Jahanafrooz, G. W. Lorimer and N. Ridley,
“The Morphology, Crystallography, and Chemistry of
Phases in As-cast Nickel-Aluminum-Bronze,” Metallur-
gical Transactions A, Vol. 13A, 1982, p. 1337.
doi:10.1007/BF02642870
[5] J. A. Wharton, R. C. Brik, G. Kear, R. J. K. Wood, K. R.
Stokes and F. C. Walsh, “The Corrosion of Nickel-Alu-
minum Bronze in Seawater,” Corrosion Science, Vol. 47,
2005, pp. 3336-3367.
do i:1 0. 10 16 / j. co rs ci.2 00 5. 05 . 05 3
[6] J. A. Wharton and K. R. Stokes, “The Influence of
Nickel-Aluminum Bronze Microstructure and Crevice
Solution on the Initiation of Crevice Corrosion,” Elec-
trochimica Acta, Vol. 53, 2008, pp. 2463-2473.
doi:10.1016/j.electacta.2007.10.047
[7] M. D. Fuller, S. Swaminathan, A. P. Zhilyaev and T. R.
Mcnelley, “Microstructural Transformations and Me-
chanical Properties of Cast NiAl Bronze: Effects of Fu-
sion Welding and Friction Stir Processing,” Materials
Science and Engineering A, Vol. 463, 2007, pp. 128-137.
doi:10.1016/j.msea.2006.07.157
[8] H. S. Campbell, “Aluminium Bronze Corrosion Resis-
tance Guide,” Publication 80, Copper Development As-
sociation, UK, July 1981, pp. 1-27.
[9] Z. Charlws and J. Ferrara Robert, “Sea Water Corrosion
of Nickel-Aluminum Bronze,” Transactions of the
American Foundrymen’s Society, Vol. 82, 1974, pp.
71-78.
[10] R.-P. Chen, Z.-Q. Liang, W. W. Zhang, D.-T. Zhang,
Z.-Q. Luo and Y.-Y. Li, “Effect of Heat Treatment on
Microstructure and Properties of Hot-Extruded Nic-
kel-Aluminum Bronze,” Transactions of Nonferrous
Metals Society of China, Vol. 17, 2007, pp. 1254-1258.
doi:10.1016/S1003-6326(07)60258-1
[11] A. Schussler and H. E. Exner, “The Corrosion of
Nickel-Aluminium Bronzes in Seawater-I. Protective
Layer Formation and the Passivation Mechanism,” Cor-
rosion Science, Vol. 34, 1993, p. 1793.
doi:10.1016/0010-938X(93)90017-B
[12] H. Meigh, “Cast and Wrought Aluminium Bronzes-Prop-
erties,” Processes and Structure, 1st Edition, IOM Com-
munications, 2000.
[13] F. L. LaQue, “Marine Corrosion,” Wiley, New York,
1975.
[14] J. C. Rowlands, “Studies of the Preferential Phase Corro-
sion of Cast Nickel Aluminium Bronze in Seawater,”
Proceeding of 8th International Congress of Metallic
Corrosion , 1981, p. 1346.
[15] G. Kear, B. D. Barker, K. R. Stokes and F. C. Walsh,
“Electrochemical Corrosion of Unalloyed Copper in
Chloride Media—A Critical Review,” Corrosion Science,
Vol. 47, 2004, p. 1694. doi:10.1016/j.corsci.2004.08.013
[16] G. Kear, B. D. Barker, K. R. Stokes and F. C. Walsh,
“Electrochemical Corrosion Behaviour of 90-10Cu-Ni
Alloy in Chloride-Based Electrolytes,” Journal of Applied
Electroch emist ry, Vol. 34, 2004, p. 659.
doi:10.1023/B:JACH.0000031164.32520.58
[17] G. Kear, B. D. Barker and F. C. Walsh, “Electrochemistry
of Non-Aged 90-10 Copper-Nickel Alloy (UNS C70610)
as a Function of Fluid Flow Part 1: Cathodic and Anodic
Characteristics,” Electrochimica Acta, Vol. 52, No. 5,
2007, pp. 1889-1898. doi:10.1016/j.electacta.2006.07.054
[18] K. Habib, “Measurement of the a.c. Impedance of Alu-
minum Samples by Holographic Interferometry,” Optics
and Lasers in Engineering, Vol. 28, 1997, pp. 37-46.
doi:10.1016/S0143-8166(96)00058-9
[19] K. Habib, “Zero Resistance Ammeter of Metallic Alloys
in Aqueous Solutions,” Optik-International Journal for
Light and Electron Optics, Vol. 118, No. 6, 2007, pp.
296-301. doi:10.1016/j.ijleo.2006.03.023
[20] B. Zhao, J.-H. Li, R.-G. Hu, R.-G. Du and C.-J. Lin,
“Study on the Corrosion Behavior of Reinforcing Steel in
Cement Mortar by Electrochemical Noise Measure-
ments,” Electrochimica Acta, 2007, Vol. 52, pp.
3976-3984. doi:10.1016/j.electacta.2006.11.015
[21] R. A. Cottis, “Interpretation of Electrochemical Noise
Data,” NACE International Corrosion, Vol. 57, No. 3,
2009, pp. 65-23.
[22] H.-H. Huang, W.-T. Tsai and J.-T. Lee, “The Influences
of Microstructure and Composition on the Electrochemi-
cal Behavior of a516 Steel Weldment,” Corrosion Sci-
ence, Vol. 36, No. 6, 1994, pp. 1027-1038.
doi:10.1016/0010-938X(94)90201-1
[23] F. T. Cheng, K. H. Lo and H. C. Man, “An Electro-
chemical Study of the Crevice Corrosion Resistance of
NiTi in Hanks’ Solution,” Journal of Alloys and Com-
pounds, Vol. 437, 2007, pp. 322-328.
doi:10.1016/j.jallcom.2006.07.127
[24] E. A. Culpan and R. Rose, “Corrosion Behaviour of Cast
Nickel Aluminum Bronze in Sea Water,” British Corro-
sion Journal, Vol. 14, No. 3, 1979, p. 160.
[25] G. W. Loriner, F. Hasan, J. Iqbal and N. Ridley, “Obser-
vation of Microstructure and Corrosion Behaviour of
Some Aluminum Bronzes,” British Corrosion Journal,
Vol. 21, 1986, No. 4, pp. 244-247.
[26] A. Al-Hashem and W. Riad, “The Role of Microstructure
of Nickel-Aluminium-Bronze Alloy on its Cavitation
Corrosion Behavior in Natural Seawater,” Materials
Characterization, Vol. 48, No. 1, 2002, pp. 37-41.
doi:10.1016/S1044-5803(02)00196-1
[27] H. Jafari, M. Hasbullah Idris, A. Ourdjini, H. Rahimi and
B. Ghobadian, “EIS Study of Corrosion Behavior of Me-
tallic Materials in Ethanol Blended Gasoline Containing
Water as a Contaminant,” Fuel, Vol. 90, No. 3, 2011, pp.
1181-1187. doi:10.1016/j.fuel.2010.12.010
[28] W. R. Osório, L. C. Peixoto, M. V. Canté and A. Garcia,
“Electrochemical Corrosion Characterization of Al-Ni
Alloys in a Dilute Sodium Chloride Solution,” Electro-
Copyright © 2011 SciRes. MSA
Corrosion Behavior of Heat Treated Nickel-Aluminum Bronze Alloy in Artificial Seawater
Copyright © 2011 SciRes. MSA
1555
chim Acta, Vol. 55, No. 13, 2010, pp. 4078-4085.
doi:10.1016/j.electacta.2010.02.029
[29] W. R. Osório, L. C. Peixo, D. J. Moutinho, L. G. Gomes,
I. L. Ferreira and A. Garcia, “Corrosion Resistance of
Directionally Solidified Al-6Cu-1Si and Al-8Cu-3Si Al-
loys Castings,” Materials and Design, Vol. 32, No. 7,
2011, pp. 3832-3837.
[30] W. R. Osório, L. C. Peixoto, M. V. Canté and A. Garcia,
“Microstructure Features Affecting Mechanical Properties
and Corrosion Behavior of a Hypoeutectic Al-Ni Alloy,”
Materials and Design, Vol. 31, No. 9, 2010, pp.
4485-4489. doi:10.1016/j.matdes.2010.04.045
[31] W. R. Osório, D. M. Rosa, L. C. Peixoto and A. Garcia,
“Cell/Dendrite Transition and Electrochemical Corrosion
of Pb-Sb Alloys for Lead-Acid Battery Applications,”
Journal of Power Sources, Vol. 196, No. 15, 2011, pp.
6567-6572. doi:10.1016/j.jpowsour.2011.03.050