Corrosion Science 77 (2013) 143–150

Contents lists available at ScienceDirect

Corrosion Science
journal homepage: www.elsevier.com/locate/corsci

Impurity control and corrosion resistance of magnesium–aluminum

alloy
Ming Liu a,⇑, Guang-Ling Song b,1,2
a
China Science Lab, General Motors China, No. 56 Jinwan Road, Pudong, Shanghai 201206, China
b
Materials Science and Technology Division, Oak Ridge National Laboratory, One Bethel Valley Rd., P.O. Box 2008, MS-6156, Oak Ridge, TN 37831-6165, USA

a r t i c l e i n f o a b s t r a c t

Article history: In the present study, the corrosion behavior of AXJ530 magnesium alloy with different iron and manga-
Received 18 February 2013 nese contents is investigated in 3.5 wt% sodium chloride solution in order to tailor the tolerance limit of
Accepted 25 July 2013 Fe impurity in the magnesium alloy. Through a comprehensive phase diagram calculation and corrosion
Available online 2 August 2013
evaluation, the mechanisms for the tolerance limit of Fe in magnesium alloys are discussed. The study
adds a new dimension to controlling the Mg alloy impurity in terms of alloying composition design
Keywords: and casting conditions.
A. Alloy
Ó 2013 Elsevier Ltd. All rights reserved.
A. Magnesium
A. Iron

1. Introduction
Magnesium alloys are promising materials in the automobile
and aerospace industries due to their low density and high
strength-to-weight ratio. However, their poor corrosion perfor-
mance [1–10] limits their practical applications.
Magnesium is extremely sensitive to impurity elements: Fe, Ni,
Cu and Co [1,11,12] in terms of the corrosion performance. There is
a tolerance limit for each of the impurities in magnesium or its al-
loy. The corrosion of a magnesium alloy is usually insignificant if
the concentrations of these impurities are below their tolerance
limits, whereas the corrosion rates will substantially increase after
the impurity concentrations exceed the tolerance limits, such an
increase can sometimes even exceed 100 fold, e.g., in a non-linear
fashion. Liu et al. [13] summarized the corrosion rates of various
magnesium alloys immersed in salt water. After a comparison of
these alloys’ chemical compositions, they found that the corrosion
rates of these magnesium alloys were typically below 10 mm/y
(millimeter per year); however, if the magnesium and magnesium
alloys were contaminated with Fe impurity (usually above 40
weight ppm), the corrosion rates were usually above 20 mm/y,
and in some cases, greater than 200 mm/y. Therefore, a very simple
surface cleaning process can efficiently remove the Fe contamina-
tion on a commercial magnesium alloy and lead to a dramatically
⇑ Corresponding author. Tel.: +86 21 2898 7134.
E-mail addresses: ming.liu@gm.com (M. Liu), guangling.song@gmail.com, song-
g@ornl.gov (G.-L. Song).
1
Tel.: +1 865 574 4451.
2 in different magnesium alloys? Third, what metallurgical factors
This submission was written by the author acting in his own independent
capacity and not on behalf of UT-Battelle, LLC, or its affiliates or successors.
decreased corrosion rate [14–17]. The impurity tolerance limits are
different for different impurities. The tolerance limits of Fe, Ni and
Cu are 170 weight ppm, 5 weight ppm and 1000 weight ppm,
respectively [11,12].
Other elements have an influence on the tolerance limit of Fe.
For example, aluminum alloying decreases the iron tolerance limit,
but manganese addition can increase it [13]. Riechek, Hillis, Mer-
cer, et al. studied the tolerance limits of impurities in magnesium
alloys AZ91 [18], AM60 [19], AS41 [20] and AE42 [21]. Their results
indicated that the tolerance limits of Ni and Cu are independent of
the Mn content, but the Fe tolerance limit is closely related to the
Mn content, and the corrosion rate is dependent on the weight ra-
tio of the Fe content to the Mn content. Their experimental results
revealed that there is a critical Fe to Mn weight ratio, above which
the corrosion rate of magnesium alloy sharply increases as the Fe/
Mn wright ratio increase. The critical Fe/Mn weight ratio ranges
from 0.010 to 0.032, depending on the alloy type [18–21]. These
critical Fe/Mn weight ratios have now been widely cited as an
ASTM standard [22], although there is one report [23] that states
the critical Fe/Mn weight ratios in the AM50 magnesium alloy
can be set between 0.0072 and 0.0101 while in the AM60 magne-
sium alloy it can be between 0.007 and 0.232.
Although the Fe/Mn weight ratio criterion works very well in
industrial corrosion control practice, there are a few questions
about this interesting phenomenon from a scientific point of view.
First, why is there a sudden change when the Fe/Mn weight ratio
reaches a certain value? Hillis et al. offered several possible expla-
nations [19,20], but none of them have been experimentally
proved. Second, why is this critical Fe/Mn weight ratio different
determine this ratio value?

0010-938X/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.corsci.2013.07.037
144 M. Liu, G.-L. Song / Corrosion Science 77 (2013) 143–150

These questions have not been well answered until recently Liu the influence of iron impurities on its corrosion performance has
et al. [13] tried to explain the Fe/Mn weight ratio based on calcu- not been systematically investigated and is not yet well
lated Mg–Fe phase diagrams using Pandat software, which is a understood.
software package for calculation of phase diagrams and thermody- For a wider application of magnesium alloys, it is important to
namic properties of multi-component alloys [13]. The calculated understand the mechanism of the tolerance limit of impurity. In
Mg–Fe binary phase diagram indicated that during non-equilib- this paper, several AXJ530 magnesium alloy samples were die cast
rium cooling (casting solidification), the magnesium melt contain- and a series of phase diagrams were calculated. By comparing the
ing less than 180 weight ppm Fe would solidify to a single a-Mg experimental results and the theoretic calculations, the relation-
phase with Fe in a supersaturated solid solution in the magnesium ship between the Fe impurity level and corrosion performance
lattice. A magnesium melt or casting containing more than 180 was studied and the mechanism for the tolerance limit of the Fe
weight ppm Fe would produce a separate iron-rich phase during impurity in this alloy was discussed.
cooling and solidification [13]. If this phase is an efficient cathode,
it will accelerate the alloy corrosion due to its micro-galvanic cou-
2. Experimental
pling to the magnesium matrix. In this case, the tolerance limit of
Fe in pure magnesium corresponds to the maximum content of Fe
2.1. Materials and casting
dissolved in cast magnesium above which a Fe-rich phase will pre-
cipitate from the melt before final solidification. Therefore, for cast
The commercial magnesium alloy AM50 (Mg–5 wt% Al–0.4 wt%
pure magnesium or non-equilibrium magnesium, the tolerance
Mn) was used as the base alloy. The primary AXJ530 magnesium
limit of Fe is predicted to be 180 weight ppm, which is in agree-
alloy ingot, designated as Mg–5 wt% Al–3 wt% Ca–0.15 wt% Sr
ment with the 170 weight ppm stated in literature [11,12]. In the
was prepared in a steel crucible by adding pure calcium (99 wt%
case of Mg–Al–Fe–Mn melts, the calculation for the Mg–Al–Mn–
purity) and pure strontium (99 wt% purity) to the AM50 magne-
Fe system was agreed with previous studies [18–20]. However,
sium alloy melt, which was gravity cast in a steel die and air
the calculated Fe–Mn curve gives a slope of 0.143 in the isothermal
cooled. The chemical composition of the primary AXJ530 magne-
section at 651 °C from the Mg–5 wt% Al–Mn–Fe phase diagram.
sium alloy ingot for this study is listed in Table 1.
This is much higher than the 0.01–0.032 values stated in the liter-
Alloying was performed in a 150 kg capacity induction furnace
ature [18–20].
in which the melt was stirred during the whole process to make
In the phase diagram calculation, both Al8Mn5 phase and Fe-
sure it was chemically uniform. Before each series of casting, about
rich phase can be obtained. The Fe rich phase is a strong cathode
100 kg of primary AXJ530 magnesium alloy ingot was heated to
to the Mg matrix and can cause high corrosion rates, while the Al8-
750 °C, at which temperature the desired quantities of the iron
Mn5 phase is a relatively passive phase. The Al8Mn5 phase precip-
impurity and manganese were added in the forms of Al–5 wt% Fe
itates as the Fe rich phase dissolves. In this process, the Fe content
ingot and anhydrous manganese chloride powder, respectively, in
may transfer into the Al8Mn5 phase and become inactive to the Mg
order to achieve a melting composition: 5 wt% aluminum, 3 wt%
matrix. Many studies [24–26] have mentioned the presence of Fe
calcium, 0.15 wt% strontium, 0.2 wt%, 0.4 wt% or 0.8 wt% manga-
in the Al–Mn intermetallic. Liu et al. even observed a tiny Fe-rich
nese and approximately 350–400 weight ppm iron. It has been dis-
area using back scatter electron imaging in Scanning Electron
covered that when manganese chloride is added to molten
Microscopy (SEM) [13]. However, no Fe was detected at all in other
magnesium alloy, it is instantaneously reduced to metallic manga-
work [27–29]. With the help of Scanning Kelvin Probe Force
nese, accompanied by the formation of magnesium chloride
Microscopy (SKPFM) techniques, the Volta potential difference of
[38,39]. In fact, the addition of anhydrous MnCl2 powder has be-
the Al8Mn5 phase relative to the magnesium matrix has been
come standard foundry practice for making magnesium alloy
scanned [15,24,30]. It was reported that the potential of Al8Mn5
[40], and this method has been successfully used in Dow chemical
is only about 200–250 mV more positive than that of the Mg–Al
[11,18–21,41]. It is expected that magnesium chlorides stays in the
matrix [15]. However, a much higher average potential value
slag after alloying. Four different casting temperatures (i.e. temper-
(250–350 mV) and standard deviation (90 mV) for this intermetal-
atures of the melted alloy in crucible right before casting) were
lic than those of Mg17Al12 (potential value 60–150 mV, standard
chosen: 730 °C, 710 °C, 680 °C and 650 °C. Once the desired tem-
deviation 20 mV) have also been reported in other studies
peratures were reached and maintained for 10–15 min for stabil-
[24,30]. This inconsistency in potential is possibly caused by a var-
ization, the melting was fed into a 280 ton magnesium cold
iation of the Fe content in the intermetallic particles. Nevertheless,
chamber die cast machine and cast into a small steel die. Three ser-
the maximum Volta potential difference between the Al8Mn5
ies of tensile specimens (15 mm in diameter and 185 mm in
phase and the magnesium matrix is around 500 mV, much smaller
length) were obtained from this die. The casting parameters for
than that between Fe and Mg, which is about 2000 mV. Therefore, a
each trial, including casting temperature and die temperature,
Fe-rich phase would be a strong cathodic phase that can signifi-
are listed in Table 2. In trial # X–Y, X represents desired Mn content
cantly deteriorate the corrosion resistance of the magnesium alloy,
level (1 represents 0.2 wt% of Mn, 2 represents 0.4 wt% of Mn and 3
rather than a Fe-free Al8Mn5 phase. For example, a recent study
represents 0.8 wt% of Mn); Y represents the casting trial in each de-
[23] showed that inclusions were present within magnesium alloys
sired Mn content level, where casting temperature and die temper-
with compositions of 40–80 wt% Al, 20–60 wt% Mn and 0–10 wt%
ature may change. The die was pre-heated at 300 °C before casting.
Fe, and corrosion preferentially occurred around inclusions with
However, for some trials, 200 °C was also used to study the effect of
high Fe content.
die temperature on the corrosion performance of casting magne-
AXJ530 is a creep-resistant magnesium alloy patented by GM,
sium alloy. Therefore the number of the ‘‘1#’’ samples was more
with a nominal composition of Mg–5 wt% Al–3 wt% Ca 0.15 wt%
than that of the ‘‘2#’’ and ‘‘3#’’ samples.
Sr [31]. The AXJ530 magnesium alloy has promising mechanical
properties, castability, and creep resistance and may have various
applications in the automotive industry [32,33]. Although the 2.2. Chemical composition analysis
allowable Fe impurity level is specified to be 40 weight ppm
[31], no reason has been given for the impurity tolerance limit in After the casting experiment, a section of a selected tensile bar
this alloy. There are only very limited studies on the corrosion was analyzed from each casting group to determine its chemical
resistance of AXJ530 alloy [34–37]. Moreover, the mechanism for composition. ICP/AES (Inductively-Coupled Plasma/Atomic Emis-
 M. Liu, G.-L. Song / Corrosion Science 77 (2013) 143–150 145

Table 1
Chemical composition of primary AXJ530 magnesium alloy ingot.

Mg (wt%) Al (wt%) Ca (wt%) Sr (wt%) Fe (wt%) Mn (wt%) Ni (wt%) Cu (wt%)

Bal 5.30 3.25 0.283 0.0078 0.236 0.0004 0.0032

Table 2
Casting process, chemical composition and corrosion result for each trial of samples.

trial# Casting temperature Die temperature Chemical Composition Corrosion rate Critical temperature
Al Ca Sr Mn Fe
°C °C wt% wt% wt% wt% ppm mm/y °C
1–0 646 ± 5 200 5.90 2.31 0.096 0.23 370 12.38 621.75
1–1 650 200 5.98 2.24 0.083 0.21 220 11.44 633.39
1–2 680 200 5.80 2.22 0.11 0.22 590 19.43 595.88
1–3 709 ± 11 200 5.10 1.84 0.093 0.18 511 25.07 595.77
1–4 731 ± 4 200 5.25 2.26 0.15 0.22 584 22.35 592.25
1–5 647.5 ± 4.5 300 5.17 2.00 0.088 0.19 430 17.12 598.05
1–6 680 ± 5 300 5.28 1.99 0.10 0.20 530 17.55 592.29
1–7 701.5 ± 1.5 300 5.29 2.46 0.094 0.19 490 18.40 590.95
1–8 728 ± 3 300 5.23 2.11 0.086 0.20 570 20.48 593.88
2–1 728.5 ± 2.5 300 6.09 2.08 0.045 0.59 140 8.18 729.49
2–2 708 ± 2 300 5.69 2.61 0.035 0.54 43 4.46 721.60
2–3 678.5 ± 3.5 300 5.68 2.03 0.033 0.43 52 5.85 705.31
2–4 652 ± 2 300 5.69 2.03 0.050 0.41 138 9.00 697.34
2–5 652 ± 2 200 5.29 2.04 0.042 0.45 59 7.38 703.37
3–1 728 ± 3 300 5.54 1.73 0.018 0.63 64 6.09 729.95
3–2 710 ± 3 300 5.84 1.84 0.020 0.66 37 6.90 737.65
3–3 679.5 ± 2.5 300 5.49 1.68 0.020 0.70 26 5.33 737.59
3–4 650 300 5.15 1.59 0.015 0.43 16 5.45 700.86

sion Spectroscopy) chemical analysis was performed to measure
the element content of Al, Ca, Sr, Mn, Fe, Ni and Cu (see Table 2).
2.3. Corrosion measurement
Corrosion rates were measured through immersion tests at
room temperature in 3.5 wt% NaCl solution for 72 h. The sample
was mounted in a cold setting resin with only 2.01 cm2 surface
area section exposed to the solution. The exposed specimen’s sur-
face was mechanically ground with 1200 grit SiC paper, polished to
0.5 lm diamond paste, washed with distilled water and dried by
clean compressed air. About 500 ml of solution was prepared for
each specimen using A.R. grade NaCl chemical and distilled water.
The solution temperature was maintained at 25 ± 2 °C by the air
conditioning of the lab during 72 h immersion tests and no deaer-
ation or stirring was applied to the solution during the experiment.
The hydrogen evolved during the corrosion experiment was col-
lected in a burette over the corroding sample using a procedure
similar to that described previously [42]. After immersion, the cor-
roded specimens were cleaned with distilled water and dried. The
corrosion morphology was examined using optical microscopy.
The samples were then immersed in a chromic acid solution
(200 g/L CrO3 + 10 g/L AgNO3) at ambient temperature for 5–
10 min to remove the corrosion products in order to measure their
corrosion weight loss.
3. Results
3.1. Chemical composition
The detailed chemical compositions of the samples used in this
study are listed in Table 2. The Mn content is well controlled at
three levels. For some samples, the Fe impurity level is higher than
that of the primary AXJ530, which indicates that the Fe comes not
only from the Al–5 wt% Fe ingot added during melting, but also
from the raw material impurity and the steel furnace. The Ca and
Sr contents are lower in the second and third series of castings,
which could be due to their burning consumption during melting
(about 24 h).
3.2. Corrosion rate measurement
Fig. 1 shows the corrosion rates measured hydrogen evolution
volume versus time for the AXJ530 samples immersed in 3.5 wt%
NaCl solution. To distinguish the 18 trials of samples in Fig 1, dif-
ferent symbol shape (square, circle, up triangle and down triangle),
line style (dash, solid and dot) and symbol style (filled and empty),
as shown in the legend, are used to represent different casting tem-
peratures, manganese contents and die temperatures, respectively.
For example, ‘‘2–1’’ and ‘‘3–1’’ have the same data point symbol,
but their line styles are different, indicating that they have the
same casting temperature, but their Mn contents are different.
Since ‘‘1–0’’ and ‘‘1–1’’ were cast in the same condition with similar
compositions as shown in Table 2, they have the same symbol and
line style in Fig 1. The first series of samples (dashed line), which
contain 0.2 wt% Mn, show much higher corrosion rates than the
samples from the second and third series (solid line and dotted
line), which have 0.4 wt% Mn and 0.8 wt% Mn, respectively. Other
parameters such as casting temperature and die temperature do
not seem to have a significant impact on the corrosion
performance.
The corrosion rates in millimeters per year (mm/y) for the sam-
ples can be calculated from their hydrogen evolution curves and
are listed in the corrosion rate column in Table 2. It can be see that
the corrosion rates of the first series of samples range from 12 to
25 mm/y, while the corrosion rates of the second and third series
of samples are in the range from 5 to 9 mm/y.
The corrosion rates of the three series of samples are also plot-
ted against Fe/Mn weight ratio in Fig. 2. This graph shows that the
corrosion rate increases as the increasing Fe/Mn ratio, but the
curve does not have significantly varied slopes as reported in the
literature [18–21].
146 M. Liu, G.-L. Song / Corrosion Science 77 (2013) 143–150

40
1-0
35 casng T 650 0.2 wt % Mn Die T 300
1-1
1-2
casng T 680 0.4 wt % Mn Die T 200
1-3
30 casng T 710 0.8 wt % Mn
1-4
casng T 730
1-5

Hydrogen(ml/cm )
25

2
1-6
1-7
20
1-8
2-1
15 2-2
2-3
10 2-4
2-5
5 3-1
3-2
0 3-3
0 10 20 30 40 50 60 70 80 3-4
time (h)

Fig. 1. The hydrogen evolution rate for each trial of casting sample in current study during 72 h of immersion corrosion test.

30 Mg - 5 wt % Al - 3 wt % Ca - 0.15 wt % Sr - 0.3 wt % Mn - Fe
800

25 0.2 wt % Mn Liquid
0.4 wt % Mn
0.8 wt % Mn 750
Liquid + BCC_B2
Corrosion Rate (mm/y)

20
Temperature ( C)

700
o

15
Liquid + BCC_B2 + Al8Mn5

10 650
Liquid + Al8Mn5

5
600

Liquid + Al8Mn5 + HCP

0
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 550
Fe/Mn weight ratio 0.000 0.005 0.010 0.015 0.020
Fe content (wt %)
Fig. 2. The corrosion rates against Fe/Mn weight ratio for each trial of casting
sample in current study.
Fig. 3. Calculated pseudo-binary section through the Mg–Al–Ca–Sr–Mn–Fe phase
diagrams at 5 wt% Al-3 wt% Ca-0.15 wt% Sr-0.3 wt% Mn (using Pandat software and
the thermodynamic database [43]).

4. Discussion
4.1. Phase diagram calculation
Mg–Al–Ca–Sr–Mn–Fe phase diagrams can be calculated using
Pandat software and the thermodynamic data in the latest PanMg
8 database [43]. To simplify the calculation and better understand
the presence of possible phases in the cast alloy samples, the major
alloying elements such as Al, Ca, Sr and Mn, as well as the Fe impu-
rity, were set as the input composition in all the calculations.
Fig. 3 presents a pseudo-binary section in the calculated Mg–
Al–Ca–Sr–Mn–Fe phase diagram at 5 wt% Al, 3 wt% Ca, 0.15 wt%
Sr and 0.3 wt% Mn, which is the typical composition proposed for
the AXJ530 alloy [36]. According to the calculation, to avoid the
formation of Fe rich phase BCC_B2, the casting temperature should
be below a critical level. This line is dependent on the Fe content;
the higher the Fe content, the lower the critical casting tempera-
ture required. Since the composition guidance for AXJ530 magne-
sium alloy allows 40 weight ppm Fe impurity [36], the phase
fractions of AXJ530 magnesium alloy containing this amount (40
weight ppm) of Fe during equilibrium solidification are calculated
and illustrated in Fig. 4. The Fe rich phase BCC_B2 disappears be-
low 670 °C and the Fe content transfers into passive phase Al8Mn5.
This indicates that, if the melt temperature is maintained below
670 °C before casting, the Fe-rich phase would disappear. However,
if the casting temperature exceed that temperature, the Fe-rich
phase would precipitate as a strong cathodic phase in the micro-
structure. According to this phase diagram calculation, the critical
casting temperature should be 670 °C, slightly lower than that
specified in the original guidance for high pressure diecast
AXJ530 magnesium alloy (which is 680 °C) [36]. An important
implication from analyzing the calculated phase diagram is that,
a small variation in composition or temperature can lead to the
precipitation of a Fe rich phase in casting samples, deteriorating
the corrosion performance of the AXJ530 magnesium alloy.
To further investigate the quantitative relationship among cast-
ing temperatures, Mn content and Fe rich phase, a series of phase
diagrams were calculated for AXJ530 magnesium alloy samples
with different Mn and Fe contents cast at different temperatures.
Fig. 5 is a calculated pseudo-binary section from the Mg–Al–Ca–
Sr–Fe–Mn phase diagram at 5 wt% Al, 3 wt% Ca, 0.15 wt% Sr and
 M. Liu, G.-L. Song / Corrosion Science 77 (2013) 143–150 147

Mg - 5 wt % Al -3 wt % Ca - 0.15 wt % Sr - 0.3 wt % Mn - 40 weight ppm Fe 750

0
10
1.0x10
-4
40 ppm Fe Liquid + BCC_B2
9.0x10
-5
100 ppm Fe
200 ppm Fe
-5
8.0x10
-1
10 7.0x10
-5 f(Al8Mn5)

-5
f(BCC_B2) 400 ppm Fe

Phase fraction
6.0x10
-5 700 1000 ppm Fe
5.0x10

Temperature( C)
-2
10 4.0x10
-5

o
Phase fractions

-5
3.0x10
-5
2.0x10

-3 -5

10 1.0x10

0.0
660 670 680 690 700 710 720 730 740 750 760 770

Temperature( C)
o
650
-4
10
f(HCP)
f(Al8Mn5) Liquid + Al8Mn5
f(Mg17Sr2)
-5
10 f(C36)
f(Liquid)
f(BCC_B2) 600
-6
10 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
500 550 600 650 700 750 800 850 900 Mn(wt %)
o
Temperature ( C)
Fig. 6. Critical casting temperature vs Mn content under different level of Fe
Fig. 4. Calculated phase fractions for equilibrium cooling of an AXJ530 Mg alloy, impurity levels.
namely 5 wt% Al, 3 wt% Ca, 0.15 wt% Sr, 0.3 wt% Mn and 40 weight ppm Fe (using
Pandat software and the thermodynamic database [43]).
casting temperature is recommended to be around 680 °C and the
Mn content about 0.3 wt%. Based on the above analysis, these spec-
Mg - 5 wt % Al - 3 wt % Ca - 0.15 wt % Sr - 200 weight ppm Fe - Mn
750 ifications are likely to lead to the precipitation of the Fe rich phase,
even though there is only 40 weight ppm of Fe in the melt. How-
ever, if the casting temperature could decrease by 30 °C to 650 °C
Liquid + BCC_B2
and the Mn content could increase by only 0.1–0.4 wt%, the sample
700 would be free of Fe rich phase according to the calculated phase
diagram.
The curves of the critical casting temperature against Fe impu-
Temperature( C)

Liquid + Al8Mn5
o

rity content at different Mn contents are calculated and plotted in

650 Fig. 7, showing a similar trend as shown in Fig. 6, that a higher Mn
content and a lower casting temperature can tolerate more Fe
Liquid + BCC_B2 + Al8Mn5 impurity in AXJ530 magnesium alloy.
It should be noted that the above analyses are simply based on
600 thermodynamic equilibrium phase diagrams. The phase transfor-
Liquid + BCC_B2 + HCP + Al8Mn5
mation during a practical casting or solidification process is much
more complicated. In theory, in the BCC region of the phase dia-
Al8Mn5 + Liquid + HCP
gram, the Fe rich phase can precipitate from the melt. In reality,
550 the precipitation needs nucleation and the nucleus needs some
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
time to grow. Obviously, the precipitation is much easier from a
Mn content(wt %)

Fig. 5. Calculated pseudo-binary section through the Mg–Al–Ca–Sr–Fe–Mn phase

diagram at 5 wt% Al, 3 wt% Ca, 0.15 wt% Sr and 200 weight ppm Fe (using Pandat
750
software and the thermodynamic database [43]).
Liquid + BCC_B2

200 weight ppm Fe. With 200 weight ppm of Fe in the melt, there 700
will be a Fe rich phase precipitated in the cast sample and it will
have poor corrosion resistance if the alloy is cast above this tem-
Temperature( C)
o

perature. If the alloy is cast below that temperature, the Fe-rich

phase will disappear and the Fe will transfer into the passive Al8- 650
Mn5 phase via: Liquid + BCC_B2 ? Al8Mn5. Similarly, there is also
a critical Mn content at a certain casting temperature. The Mn con-
tent needs to exceed that critical value to avoid the precipitation of
600 0.2% Mn
Fe rich phase before casting. 0.3% Mn
The curves of the critical casting temperature per Mn content at 0.4% Mn
different levels of Fe impurity are calculated and plotted together 0.6% Mn Liquid + Al8Mn5
in Fig. 6. To improve the corrosion resistance of the AXJ530 magne- 0.8% Mn
550
sium alloy, i.e. to avoid the Fe rich phase precipitation zone in the 0.00 0.01 0.02 0.03 0.04 0.05
left top region, the Mn content needs to be high enough while the
Fe(wt%)
casting temperature needs to be sufficiently low. This actually pro-
vides a guide for casting corrosion resistant AXJ530 magnesium al- Fig. 7. Critical casting temperature vs Fe impurity content under different level of
loy. In the original AXJ530 magnesium alloy specification [36], the Mn content.
148 M. Liu, G.-L. Song / Corrosion Science 77 (2013) 143–150

melted alloy than that in a solidified one due to the much faster
diffusion of elements in the liquid phase. E.g., in a pure magnesium
system, when the content of Fe impurity is lower than 180 weight
ppm, the liquid region (HCP + liquid) above the solid region
(HCP + BCC) is very narrow (less than 0.1 °C) [13]. A Fe nucleus can-
not be formed so quickly during casting. If the content of Fe impu-
rity in pure magnesium is higher than 180 weight ppm, the liquid
region (BCC + liquid) above the solid region (HCP + BCC) is much
wider (about 10 °C for 200 weight ppm of Fe and about 50 °C for
300 weight ppm of Fe) [13]. The Fe impurity has enough time to
nucleate, grow and become a Fe rich phase during the solidifica-
tion. Similarly, in Mg–Al systems, including the AXJ530 magne-
sium alloy and other Al containing magnesium alloy in literature
[18–20], it is easy and quick for the Fe impurity to precipitate from
the liquid melt, but the transfer of Fe content from precipitated Fe
rich phase to solid Al8Mn5 intermetallic is much slower, which re-
quires the system to be held at a high temperature for a much
longer time. When AXJ530 magnesium alloy is cast from a casting
temperature below the critical temperature, it means that the alloy
has been melted and held at that casting temperature long enough
for the Fe impurity to be combined into a passive Al8Mn5 phase. If
the casting temperature is higher than the critical temperature,
due to the wide liquid region (BCC + liquid), the Fe impurity from
the raw materials and steel crucible has enough time to precipitate
and grow before getting in to the liquid region (Al8Mn5 + liquid).
After that, as the alloy system passes the liquid region (Al8Mn5 + -
liquid) quickly during the solidification process, only a very limited
Fe content can transfer from the precipitated Fe rich phase into a
passive Al8Mn5 phase. Therefore, from a kinetic point of view,
the above discussion regarding critical temperature for the toler-
ance limit of the Fe impurity still makes sense and the calculated
phase diagrams still provide a basis for prediction of corrosion per-
formance of AXJ530 magnesium alloy.
From the chemical composition obtained from ICP measure-
ments for each sample, the critical temperatures for all the trial al-
loy specimens in the current study can be calculated, and are listed
in Table 2. With these, the corrosion rates can be plotted against
the difference between casting temperature and critical tempera-
ture. Fig 8 shows that when a casting temperature is above its cor-
responding critical temperature, the Fe rich phase is present, and
the corrosion rate is higher than 10 mm/y. The corrosion rate in-
creases in an almost linear fashion up to 25 mm/y as the tempera-
ture difference increases. However, if the difference is negative,
which means that the sample was cast below that critical temper-
ature and there is no Fe rich phase precipitated in the alloy, the
corrosion rate stays relatively low, randomly between 5 and
10 mm/y. For example, in the first series of samples, which have
only 0.2 wt% Mn, there are Fe rich phase precipitates even the cast-
ing temperature is relatively low (650 °C). The second and third
series of samples, which have 0.4 wt% and 0.8 wt% Mn, respec-
tively, do not have Fe rich phase precipitated during casting. There-
fore, the corrosion rates of the first series of samples are much
higher than those of the second and third series.
The Fe rich phase has been identified under SEM using EDS
analysis in Mg–Al alloy [13]. However, for such a small amount
of Fe content (100 ppm), the Fe rich phase cannot be easily lo-
cated and identified in AXJ alloy. TEM or Atom Probe Tomography
(APT) is powerful technical method to give the phase distribution
result. However, they can only measure phases in a small area
where may not contain the Fe rich phase. It would be valuable if
further observation and analysis can be carried out to understand
the precipitation of Fe rich phase and how it transfers to the Al8-
Mn5 phase. This task is beyond the present study and will be in-
cluded in the future work.
4.2. Literature reevaluation
The work by Riechek, Hillis, Mercer, et al. [18–21] is the only
complete study regarding the dependence of corrosion resistance
of magnesium alloys on Fe/Mn weight ratio. There is no other pub-
lished result to confirm their Fe/Mn curves. In fact, the tolerance
limit of Fe/Mn weight ratio was not found from the corrosion rate
curve against Fe/Mn weight ratio directly. For example, in their
study [18–21], the Fe and Mn contents of various AM60 magne-
sium alloy samples were plotted in a chart (see Fig 9). The numbers
adjacent to the data points are salt spray corrosion rates in milli-
meter per year (mm/y) while the casting temperatures are indi-
cated in the legend; all the presented experimental results are
extracted from literature [19]. In such a matrix of Fe content (in
weight ppm) vs. Mn content (in wt%), it was found that there
was a dotted line starting from the origin point which divided
these points into two groups. In the upper left region, the points
have very high corrosion rates, from 0.76 to 11.9 mm/y, while in
the lower right region, the points have relatively low corrosion
rates, from 0.23 to 0.46 mm/y. The slope of this dotted line repre-
sents the critical Fe/Mn ratio. After that, a diagram was plotted

250
30

200
11.9
25 0.2 wt % Mn
0.4 wt % Mn 97
9.7
Corrosion Rate(mm/y)

0.8 wt % Mn
Fe(weightt ppm)

20 150
3.8
10 7
10.7 59
5.9
15 Experiment at 650
100 Experiment at 680
5.3 Experiment at 720
2.1 0.46
10 Experiment
pe e a at 750
50
Calculation at 650
50 0.76 Calculation at 680
0.46 0.23
5 Calculation at 720
0.23 Calculation at 750
Experiment Prediction
0
0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
-100 -50 0 50 100 150
Mn(wt %)
o
Difference between Casting T and Critical T( C)
Fig. 9. The experiment result and phase diagram calculation of AM60 sample cast
Fig. 8. The corrosion rates vs the difference between casting temperature and at different temperature [19]. The numbers adjacent to data points are salt spray
critical temperature for each trial of casting sample in current study. corrosion rates in mm/y.
 M. Liu, G.-L. Song / Corrosion Science 77 (2013) 143–150
with a corrosion rate against Fe/Mn weight ratio curve to refine the
critical Fe/Mn weight ratio [19]. Undoubtedly, the Fe/Mn weight
ratio based on such a dotted line is quite rough and can deviate sig-
nificantly. The slope of this dotted line is 0.21, while any line with a
slope from 0.152 to 0.0274 can divide these points into two groups.
Calculated the phase diagrams offer another approach to clas-
sify the corrosion rates of Mg alloys. As the average Al content in
these three casting samples is known from literature [19], the iso-
thermal section at 650 °C, 680 °C, 720 °C and 750 °C though the
Mg–Al–Mn–Fe phase diagram can be calculated. In each calculated
Fe–Mn isothermal section, there is an isothermal line, which repre-
sents the precipitation of the Fe rich phase from the melt AM60
magnesium alloy at the specific Fe and Mn contents. These calcu-
lated isothermal lines are overlapped on the Fe–Mn matrix as
shown in Fig 9. They divide the experimental points into two
groups. For examples, for three AM60 samples cast at 680 °C, it is
clearly shown that two of the experimental points with high corro-
sion rates (5.3 and 2.1 mm/y) stay the left to the calculated 680 °C
Fe–Mn isothermal line, while the other one with a low corrosion
rate (0.23 mm/y) keeps the right to the calculated line. The isother-
mal sections at other temperatures (650 °C, 720 °C and 750 °C)
through the Mg–Al–Mn–Fe phase diagram show similar character-
istics in Fig. 9. Unfortunately, these calculated Fe–Mn isothermal
lines failed to clearly separate the corrosion rate points into two
groups as the dotted line in reference [19], as the Fe–Mn isother-
mal line shifts with temperature. It is not surprising that Liu
et al. [13] failed to match the Fe/Mn weight ratio with their calcu-
lated Fe–Mn curve at 651 °C. Nevertheless, the Fe–Mn isothermal
line is more reasonable in theory. Based on these isothermal lines
at different temperatures, it can also be concluded that a higher Mn
content or lower casting temperature is preferred to avoid Fe
impurity contamination and thus the poor corrosion performance
of a magnesium alloy.
The same validation can be done for magnesium alloys AZ91
and AS41, reported in literature [18,20] showing similar results,
except AE42, the Pandat software database does not contain rare
earth elements. The calculated isothermal Fe–Mn curves can be
used to differentiate the corrosion rates of these cast samples.
The Fe tolerance limits are better defined by the theoretical iso-
thermal Fe–Mn curve than the experimental Fe/Mn weight ratio
that is independent of liquid metal temperature. For those samples
left to the isothermal line, the farther the sample point is away
from the calculated isothermal line, the higher the corrosion rate.
In fact, the distance of the sample point to the isothermal line
reflects the difference between casting temperature and critical
temperature. To confirm this, the critical temperatures for the sam-
ples in literature [18–20] are recalculated using Pandat software.
Their reported corrosion rates [18–20] are plotted versus the dif-
ferences between their casting temperatures and calculated critical
temperatures in Fig 10. It shows that if the temperature difference
is positive, which means the sample was cast at a temperature
higher than the critical temperature and the Fe rich phase precip-
itated in the alloy, the corrosion rate increases quickly with
increasing temperature differences; if the difference is negative,
which means the sample was cast below that critical temperature
and free of Fe rich phase, the corrosion rate is very low. This trend
is very similar to that in Fig 8. Therefore, the Fe rich phase theory
can also successfully explain the experimental results of other
magnesium alloys in literature [18–20] other than AXJ530 magne-
sium alloys. The difference between casting temperature and crit-
ical temperature provides a better correlation with corrosion rates
than the Fe/Mn weight ratio.
Although the Fe rich phase theory better explains the Fe toler-
ance limit in magnesium than the Fe/Mn weight ratio, the phase
diagram calculation is simply based on the overall composition
of the alloy. It is unable to address issues associated with local met-
149
15
AZ91
AM60
AS41
Corrosion Rate(mm/y)
10
5
0
-50 0 50 100 150
O
Difference between casting T and critical T ( C)
Fig. 10. The corrosion rates against the difference between casting temperature and
critical temperature for each trial of casting sample in literature [18–20].
allurgical changes during melting and casting. For example, in lit-
erature [18–21], the Fe content decreased with lowering casting
temperature, but the slag contained a significantly higher amount
of Fe content. This difference in the Fe content between the casting
and slag cannot be reflected in the phase diagram calculation. This
means that in the phase diagram calculation, the change of casting
temperature only affects the distribution of the Fe content in the Fe
rich phase or magnesium solid solute matrix, and its influence on
the loss of the Fe content in the slag is totally ignored, which can
introduce a significant error to the calculated results.
4.3. Possible mechanism of the tolerance limit of fe in magnesium
alloys
Based on the experiment results from literature [18–21] and
present study, the possible mechanism regarding the Fe tolerance
limit in a magnesium alloy can be proposed as follows. During
the melting stage, manganese combines with a certain amount of
Fe impurity to form an intermetallic compound. Because of the
higher density of this Fe–Mn compound, it settles down to the melt
bottom to form slag. In the melted alloy liquid, the temperature
determines the form of the rest of Fe content. For certain amounts
of Mn and Fe in the melt, there is a critical temperature. Above this
temperature, the Fe content can nucleate and precipitate into a so-
lid Fe rich phase. The Fe rich phase precipitate stays in the alloy
after rapid solidification. The Fe rich precipitate with a noble corro-
sion potential leads to strong micro-galvanic corrosion of the sur-
rounding magnesium a-matrix. If the melted alloy is maintained at
a temperature lower than the critical temperature for a certain
time, the Fe content will transfer into Al8Mn5 phase in the melted
alloy liquid. The Fe content stays in the Al8Mn5 phase and is inac-
tive to the matrix phase. The Al8Mn5 has a corrosion potential clo-
ser to that of the surrounding magnesium a-matrix than the Fe rich
precipitate. As a result, the corrosion of the magnesium alloy be-
comes less severe.
5. Conclusions
1. The tolerance limit of Fe impurity in a magnesium alloy is
determined by the precipitation of the Fe rich phase.
2. The Mn content and casting temperature can influence the pre-
cipitation of the Fe rich phase and thus the corrosion perfor-
mance of a magnesium alloy in the following ways: (1) for a
150 M. Liu, G.-L. Song / Corrosion Science 77 (2013) 143–150
fixed amount of Fe contamination, a higher Mn content allows a
magnesium alloy to be cast at a higher casting temperature and
the cast alloy still has high corrosion resistance; (2) at a given
amount of Mn content, a lower casting temperature can tolerate
a higher Fe contamination for a magnesium alloy without sacri-
ficing its high corrosion resistance; and (3) under a designated
casting temperature, a higher Mn content means that a higher
Fe contamination will not lead to deteriorated corrosion resis-
tance of a magnesium alloy.
3. The quantitative relationship among the Fe contamination, Mn
content and casting temperature for a magnesium alloy can be
calculated from phase diagrams. For AXJ530 magnesium alloy,
0.4 wt% of Mn content and casting temperature of 650 °C are
recommended to reduce Fe rich phase precipitation in order
to obtain high corrosion resistance.
4. The Fe rich phase precipitation theory can more reasonably
explain the tolerance limit of Fe than the usually used Fe/Mn
weight ratio.
Acknowledgments
The author would like to thank Dr. Bin Hu and Mr. Pan Wang in
GM R&D China Science Lab for their help with casting experiments
and Dr Bob R. Powell, Dr. Alan Luo and Dr. Anil K. Sachdev in GM
R&D Warren for providing valuable discussion and suggestions.
References
[1] G. Song, A. Atrens, Corrosion mechanisms of magnesium alloys, Advanced
Engineering Materials 1 (1999) 11–33.
[2] G. Song, A. Atrens, Understanding magnesium corrosion – a framework for
improved alloy performance, Advanced Engineering Materials 5 (2003) 837–
858.
[3] G. Song, Recent progress in corrosion and protection of magnesium alloys,
Advanced Engineering Materials 7 (2005) 563–586.
[4] G. Song, A. Atrens, Recent insights into the mechanism of magnesium
corrosion and research suggestions, Advanced Engineering Materials 9
(2007) 177–183.
[5] X.B. Chen, N. Birbilis, T.B. Abbott, Review of corrosion-resistant conversion
coatings for magnesium and its alloys, Corrosion 67 (2011).
[6] X.B. Chen, H.Y. Yang, T.B. Abbott, M.A. Easton, N. Birbilis, Corrosion-resistant
electrochemical platings on magnesium alloys: a state-of-the-art review,
Corrosion 68 (2012) 518–535.
[7] N.T. Kirkland, N. Birbilis, M.P. Staiger, Assessing the corrosion of biodegradable
magnesium implants: a critical review of current methodologies and their
limitations, Acta Biomaterialia 8 (2012) 925–936.
[8] N.T. Kirkland, G. Williams, N. Birbilis, Observations of the galvanostatic
dissolution of pure magnesium, Corrosion Science 65 (2012) 5–9.
[9] R. Zeng, J. Zhang, W. Huang, W. Dietzel, K.U. Kainer, C. Blawert, W. Ke, Review
of studies on corrosion of magnesium alloys, Transactions of Nonferrous
Metals Society of China 16 (2006) s763–s771.
[10] E. Ghali, W. Dietzel, K.-U. Kainer, General and localized corrosion of
magnesium alloys: a critical review, Journal of Materials Engineering and
Performance 13 (2004) 7–23.
[11] J.D. Hanawalt, C.E. Nelson, J.A. Peloubet, Corrosion studies of magnesium and
its alloys, Transactions of the American Institute of Mining and Metallurgical
Engineers 147 (1942) 273–298.
[12] G.L. Makar, J. Kruger, Corrosion of Magnesium, International Materials Reviews
38 (1993) 138–153.
[13] M. Liu, P.J. Uggowitzer, A.V. Nagasekhar, P. Schmutz, M. Easton, G.-L. Song, A.
Atrens, Calculated phase diagrams and the corrosion of die-cast Mg–Al alloys,
Corrosion Science 51 (2009) 602–619.
[14] G.-L. Song, Z. Xu, The surface, microstructure and corrosion of magnesium
alloy AZ31 sheet, Electrochimica Acta 55 (2010) 4148–4161.
[15] G.-L. Song, Z.Q. Xu, Effect of microstructure evolution on corrosion of different
crystal surfaces of AZ31 Mg alloy in a chloride containing solution, Corrosion
Science 54 (2012) 97–105.
[16] U.C. Nwaogu, C. Blawert, N. Scharnagl, W. Dietzel, K.U. Kainer, Influence of
inorganic acid pickling on the corrosion resistance of magnesium alloy AZ31
sheet, Corrosion Science 51 (2009) 2544–2556.
[17] U.C. Nwaogu, C. Blawert, N. Scharnagl, W. Dietzel, K.U. Kainer, Effects of
organic acid pickling on the corrosion resistance of magnesium alloy AZ31
sheet, Corrosion Science 52 (2010) 2143–2154.
[18] K.N. Riechek, K.J. Clark, J.E. Hillis, Controlling the Salt Water Corrosion
Performance of Magnesium AZ91 Alloy, Society of Automotive Engineers
Technical Paper 850417, Inc., Warrendale, Pennsylvania, USA, 1985.
[19] J.E. Hillis, K.N. Reichek, High Purity Magnesium AM60 Alloy: The Critical
Contaminant Limits and the Salt Water Corrosion Performance, Society of
Automotive Engineers Technical Paper 860288, Inc., Warrendale,
Pennsylvania, USA, 1986.
[20] J.E. Hillis, S.O. Shook, Composition and Performance of an Improved
Magnesium AS41 Alloy, Society of Automotive Engineers Technical Paper
890205, Inc., Warrendale, Pennsylvania, USA, 1989.
[21] W.E. Mercer, J.E. Hillis, The Critical Contaminant Limits and Salt Water
Corrosion Performance of Magnesium AE42 Alloy, Society of Automotive
Engineers Technical Paper 920073, Inc., Warrendale, Pennsylvania, USA, 1992.
[22] ASTM-B94, Standard specification for magnesium-alloy die castings, 2007.
[23] H. Matsubara, Y. Ichige, K. Fujita, H. Nishiyama, K. Hodouchi, Effect of impurity
Fe on corrosion behavior of AM50 and AM60 magnesium alloys, Corrosion
Science 66 (2013) 203–210.
[24] F. Andreatta, I. Apachitei, A.A. Kodentsov, J. Dzwonczyk, J. Duszczyk, Volta
potential of second phase particles in extruded AZ80 magnesium alloy,
Electrochimica Acta 51 (2006) 3551–3557.
[25] O. Holta, H. Westengen, J. Roen, High purity magnesium die casting alloys:
impact of metallurgical principles on industrial practice, in: Proceedings of the
Third International Magnesium Confernece, Manchester, UK, 1996.
[26] J.-Y. Byun, S. Kwon, H. Ha, J.-K. Yoon, A manufacturing technology of AZ91-
alloy slurry for semi solid forming, in: Proceedings of the 6th International
Conference Magnesium Alloys and their Applications, Wolfburg, Germany,
2003.
[27] S. Lun Sin, D. Dubé, R. Tremblay, Characterization of Al–Mn particles in AZ91D
investment castings, Materials Characterization 58 (2007) 989–996.
[28] R.M. Wang, A. Eliezer, E. Gutman, Microstructures and dislocations in the
stressed AZ91D magnesium alloys, Materials Science and Engineering: A 344
(2003) 279–287.
[29] V.Y. Gertsman, J. Li, S. Xu, J.P. Thomson, M. Sahoo, Microstructure and second-
phase particles in low- and high-pressure die-cast magnesium alloy AM50,
Metallurgical and Materials Transactions A – Physical Metallurgy and
Materials Science 36A (2005) 1989–1997.
[30] M. Ben-Haroush, G. Ben-Hamu, D. Eliezer, L. Wagner, The relation between
microstructure and corrosion behavior of AZ80 Mg alloy following different
extrusion temperatures, Corrosion Science 50 (2008) 1766–1778.
[31] B. Powell, V. Rezhets, A. Luo, B. Tiwari, Creep-resistant magnesium alloy die
castings, US Patent 6264763 B1.
[32] B.R. Powell, A.A. Luo, V. Rezhets, J.J. Bommarito, B.L. Tiwari, Development of
Creep-Resistant Magnesium Alloys for Powertrain Applications: Part 1 of 2,
Society of Automotive Engineers Technical Paper 2001-01-0422, Inc.,
Warrendale, Pennsylvania, USA, 2001.
[33] A.A. Luo, M.P. Balogh, B.R. Powell, Tensile Creep and Microstructure of
Magnesium–Aluminum–Calcium-Based Alloys for Powertrain Applications –
Part 2 of 2, Society of Automotive Engineers Technical Paper 2001-01-0423,
Inc., Warrendale, Pennsylvania, USA, 2001.
[34] D. Bharadwaj, Y. Wang, Die cast magnesium components, US Patent
20090071620 A1.
[35] G.-L. Song, D. Radovic, Lab Evaluation and Comparison of Corrosion
Performance of Mg Alloys, Society of Automotive Engineers Technical Paper
2010-01-0728, Warrendale, Pennsylvania, USA, 2010.
[36] R. Beals, Z.-K. Liu, J. Wayne Jones, P. Mallick, D. Emadi, D. Schwam, B. Powell,
USAMP magnesium powertrain cast components: fundamental research
summary, JOM Journal of the Minerals, Metals and Materials Society 59
(2007) 43–48.
[37] D. Sivaraj, R.C. McCune, P.K. Mallick, P. Mohanty, Aqueous Corrosion of
Experimental Creep-Resistant Magnesium Alloys, Society of Automotive
Engineers Technical Paper 2006-01-0257, Inc., Warrendale, Pennsylvania,
USA, 2006.
[38] W.R. Veazey, Manganese–magnesium alloy and method of making same, US
Patent 1377374.
[39] C.E. Nelson, G.F. Gunn, Method of alloying magnesium with manganese, US
Patent 2261906.
[40] E.F. Emley, Principles of Magnesium Technology, Pergamon Press, New York,
1966. pp. 670–735.
[41] J.E. Hillis, The Effects of Heavy Metal Contamination on Magnesium Corrosion
Performance, Society of Automotive Engineers, Inc., Warrendale, Pennsylvania,
USA, 1983.
[42] G.L. Song, A. Atrens, D. StJohn, in: J.N. Hryn (Ed.) Magnesium Technology 2001
Symposium, Minerals, Metals & Materials Soc, New Orleans, La, 2001, pp. 255–
262.
[43] Pandat is a software package for calculating phase diagrams and
theromdynamic properties of multi-component alloys, CompuTherm LLC,
437 S. Yellowstone Dr. Suite 217, Madison, WI 53719 USA, Phone: +1 (608)
274–1414, fax: +1 (608) 274–6045, Email: info@computherm.com.