Light metal compound casting

J.M.PAPIS Konrad, F.LOEFFLER Joerg and J.UGGOWITZER Peter

Citation: Science in China Series E-Technological Sciences 52, 46 (2009); doi: 10.1007/s11431-008-0345-9

View online: http://engine.scichina.com/doi/10.1007/s11431-008-0345-9

View Table of Contents: http://engine.scichina.com/publisher/scp/journal/Sci China Tech Sci E/52/1

Published by the Science China Press

Articles you may be interested in

Understanding of the influence of process parameters on the heat transfer behavior at the metal/die interface in high pressure die casting
process
Science in China Series E-Technological Sciences 52, 172 (2009);

ULTRAVIOLET SENSITIVITY PEAK OF INSECT COMPOUND EYE SHIFTING WITH LIGHT INTENSITY
Chinese Science Bulletin 33, 1034 (1988);

CALCULATIONS FOR CRYSTAL COHESIVE ENERGY OF TRANSITION METAL COMPOUND

Science in China Series A-Mathematics, Physics, Astronomy & Technological Science 32, 351 (1989);

Preparation of Metal Particle Catalysts Wrapped in Organosilicon Compound via Metal Vapor Synthesis
Science in China Series B-Chemistry, Life Sciences & Earth Sciences 36, 1153 (1993);

Reversible switching between normal and thermally activated delayed fluorescence towards “smart” and single compound white-light
luminescence via controllable conformational distribution
SCIENCE CHINA Chemistry 61, 677 (2018);
Science in China Series E: Technological Sciences www.scichina.com
© 2009 SCIENCE IN CHINA PRESS tech.scichina.com
www.springerlink.com
Springer

Light metal compound casting

Konrad J. M. PAPIS†, Joerg F. LOEFFLER & Peter J. UGGOWITZER
ETH Zurich, Laboratory of Metal Physics and Technology, Wolfgang-Pauli-Str. 10, 8093 Zurich, Switzerland

‘Compound casting’ simplifies joining processes by directly casting a metallic melt onto a solid metal
substrate. A continuously metallurgic transition is very important for industrial applications, such as
joint structures of spaceframe constructions in transport industry. In this project, ‘compound casting’
of light metals is investigated, aiming at weight-saving. The substrate used is a wrought aluminium
alloy of type AA5xxx, containing magnesium as main alloying element. The melts are aluminium alloys,
containing various alloying elements (Cu, Si, Zn), and magnesium. By replacing the natural oxygen
layer with a zinc layer, the inherent wetting difficulties were avoided, and compounds with flawless
interfaces were successfully produced (no contraction defects, cracks or oxides). Electron microscopy
and EDX investigations as well as optical micrographs of the interfacial areas revealed their continu-
ously metallic constitution. Diffusion of alloying elements leads to heat-treatable microstructures in the
vicinity of the joining interfaces in Al–Al couples. This permits significant variability of mechanical
properties. Without significantly cutting down on wettability, the formation of low-melting intermetallic
phases (Al3Mg2 and Al12Mg17 IMPs) at the interface of Al–Mg couples was avoided by applying a protec-
tive coating to the substrate.

compound, casting, aluminium alloys, magnesium alloys, interface structure

1 Introduction
Lightweight construction in the transport industry helps
reduce weight and thus save fuel. To optimize perform-
ance, a combination of materials is the most efficient
method, because one material alone is often insufficient.
Light metals are not easy to join, though. Weak links
arise at the joints, such as rivets, welds or brazing con-
nections. Furthermore, the natural oxide layer on alu-
minium, for example, is thermodynamically stable and
inert, has a melting point much higher than the bulk ma-
terial, and is not easily wettable by metallic melts.
If a metallic melt is cast onto or around a solid metal
shape, forming a diffusion reaction zone and thus a con-
tinuous metallic transition, one talks about compound
casting[1]. A prerequisite is the formation of zones where
the cast alloy’s components diffuse into the solid mate-
persist during the casting procedure and inhibit the for-
mation of the continuous metallic junction needed for a
‘compound cast’ part.
The present work focuses on eliminating the oxide
layer and the development of a coating which will in-
hibit repassivation and thus make the aluminium sub-
strate’s surface wettable by metallic melts[2]. Light metal
couples (Al–Al and Al–Mg) were produced by com-
pound casting on laboratory scale.
2 Experimental procedures
2.1 Substrate coatings
One very efficient way to remove the natural oxide layer
from aluminium and to prevent its formation after the
treatment is the application of the so-called zincate
process[3 6]. It replaces the Al2O3 layer with a metallic
－

rial, partly via the formation of solid solutions and partly

Received June 22, 2008; accepted August 22, 2008
via the formation of intermetallic phases. If the alumin- doi: 10.1007/s11431-008-0345-9
ium oxide layer remains on the substrate’s surface it will †
Corresponding author (email: konrad.papis@mat.ethz.ch)

Sci China Ser E-Tech Sci | Jan. 2009 | vol. 52 | no. 1 | 46-51
Downloaded to IP: 192.168.0.213 On: 2019-12-10 12:42:46 http://engine.scichina.com/doi/10.1007/s11431-008-0345-9
Zn film with a thickness of several hundred nanometers,
via two parallel chemical reactions[2,4,7]. First, the oxide
layer is removed by etching, while metallic Al oxidizes
and dissolves, Zn anions are reduced and deposit as a
dense metallic layer in the second reaction[8]. The solu-
tion used in this work was a single solution containing
NaOH and Zn anions, and was applied in ambient con-
ditions.
Prior to immersion in this zincate solution, the sub-
strate had to be pre-treated. As the substrate was re-
ceived in rolled condition, the outermost microns (‘roll-
ing layers’) had to be removed. The samples were thus
cleaned in a mild pickle solution at 55℃ to remove car-
bon contaminations, and subsequently the aluminium
oxide layer was etched away in a strong alkaline NaOH
solution (pH > 13) for 2 min at 55℃. The pickling layer,
which was built up during this process, was removed by
etching in 35% nitric acid. Between each of these im-
portant cleaning processes, the samples were rinsed in
deionized water for 2 min.
This zincate treatment removed the aluminium oxide
layer permanently, but did not yet yield the good wetting
properties needed for successful compound casting. For
Al–Al compound casting, the coating thickness was in-
creased to 10 μm via zinc galvanizing. For Al–Mg
compounds, another coating had to be applied to prevent
the excessive formation of brittle intermetallic phases.
Manganese was chosen for this purpose, and a 5-μm
thick Mn layer was deposited via galvanizing[9].
2.2 Wetting experiments
Aluminium-based alloy AlMg1 (AA5005; 1.0 Mg, 0.4
Fe, 0.25 Si, 0.15 Mn――all in mass-%) platelets of
20 mm × 10 mm × 2 mm were used as substrate for wet-
ting experiments. The melts for Al–Al compounds were
pure 99.98%-aluminium (Al), aluminium alloys with
7 mass-% Si (AlSi7), 7 mass-% Cu (AlCu7) and
7 mass-% Zn (AlZn7), respectively. For Al–Mg, only
pure 99.98%-magnesium was used.
The setup for wetting experiments comprises a hori-
zontal furnace with a device designed both to drop me-
tallic melts onto a substrate and to break and shear the
Table 1 Parameters for sample production and heat-treatment procedures
Cast alloy Solution annealing (only for T6) Ageing temp. (℃)
AlCu7 190
500℃
AlSi7 170
0.5 h
AlZn7 120
air
Al 170a)
a) These numbers were taken to be comparable to at least one of the other specimens.
Konrad J.M. PAPIS et al. Sci China Ser E-Tech Sci | Jan. 2009 | vol. 52 | no. 1 | 46-51
Downloaded to IP: 192.168.0.213 On: 2019-12-10 12:42:46 http://engine.scichina.com/doi/10.1007/s11431-008-0345-9
oxide hull off the molten metal. For this purpose, a
quartz glass tube with a diameter of 6 mm was narrowed
at one end. 300 mg of the alloy was melted inside this
tube, and pushed out using a plunger. Upon dropping
onto the substrate’s surface, none of the oxide layer de-
bris from the melt was present at the interface, and the
sample was immediately pulled out of the heated area to
cool down. The atmosphere was Ar6 (99.9999% Ar, with
previous evacuation to at least 5×10−5 mbar), with a fur-
nace temperature of 700℃. This setup was used for all
the materials described here.
2.3 Heat treatment procedures and diffusion simu-
lations for Al–Al couples
To determine diffusion zone dimensions and the degree
of interdiffusion of alloying elements experimentally,
Al–Al couples were heat treated and the interface’s me-
chanical properties were investigated. According to the
industry heat treatments, T5 and T6 states (without and
with solution annealing before artificial ageing, respec-
tively) were established for the interface compositions,
where heat treatable compositions formed after sample
production. In Table 1, these heat treatments are listed.
The ageing temperatures and times varied for each cou-
ple composition according to treatments for commercial
alloys with similar compositions[10].
Diffusion zone extensions, and thus the area of ex-
pected hardness increase, were verified by one-dimen-
sional diffusion simulations. These were performed us-
ing DICTRA software[11], with thermodynamic data
taken from COST II and aluminium alloys’ mobility
databases[12,13]. For solidification simulations, starting
temperatures were chosen to lie below the solidus of the
substrate and above the liquidus of the melts. It was not
possible to perform simulations with AlMg1/Al99.98%
couples, as the substrate’s solidus is lower than the liq-
uidus of Al. A cooling rate of 1 K/s was used for casting
simulations, and solution annealing simulations were
performed with 500℃ for 30 min. A planar solidifica-
tion front, no significant influence of convection and the
presence of only fcc and liquid phases were the simpli-
fying assumptions made.
Ageing steps (aged 1―aged 5) (h) Ageing time for T6 (expected) (h)
3, 6, 9, 12, 15 9
3, 6, 9, 12, 15 9
8, 16, 24, 32, 40 24
3, 6, 9, 12, 15 9a)
47
3 Results and discussion
3.1 Al–Al: Wetting experiments, diffusion and me-
chanical properties
AlMg1 showed a much improved wettability when zin-
cate treated and Zn galvanized. Flawless interfaces were
obtained (no pores, oxide inclusions, or cracks), as
shown in Figure 1. Narrow transitions from substrate to
the melt was observed in every sample, with a little
second phase showing up in AlMg1/AlZn7 samples. The
reference AlMg1/Al couples show a 150 μm wide zone
where the substrate was melted by the pure Al droplet
during the casting process.
Alloying elements of the substrate (Mg) and the melts
(Cu, Si, Zn) can form precipitates in the diffusion zones.
Hardness measurements can reveal diffusion zone
widths after artificial ageing, which can be verified by
subsequent diffusion simulations using DICTRA software.
AlMg1/AlSi7 couples exhibit the widest diffusion zone,
and AlMg1/AlCu7 the narrowest. Both show concentra-
tion gradients in the interfacial region, which can be
seen as continuous transition from the single-phase sub-
strate to the partially eutectic droplets in optical micro-
graphs (Figures 1(a)－1(d)). The second phase directly at
the interface in AlMg1/AlZn7 couples (Figures 1(e)－
1(f)) arises from the presence of 2.5 mass-% Zn and 0.5
mass-% Mg, whereas the Mg concentration in the bulk
is too low to form precipitates. Figures 1(g) and 1(h)
show the reference couples’ interface, where the latent
heat release at 660℃ causes a partial (150 μm deep)
melting of the substrate’s surface, which re-solidifies the
forming eutectic intermetallics with the elements present
in this commercial AA5005 alloy.
In Table 2, the concentration limits of alloying ele-
ments for commercial age-hardenable aluminium alloys
are listed. With these required minimum element con-
centrations, hardening precipitates can form. These are
β′ (Mg2Si), S′ (Al2CuMg) and η′ (Zn2Mg)[14], for the
corresponding couple compositions.
The concentration profiles in Figure 2 were obtained
from DICTRA simulations. The as-cast state is repre-
sented by the dashed curve, and the solid curves show
the composition profiles after solution annealing. Inter-
diffusion of alloying elements occurs in the vicinity of
the interface. During solidification, Si and Mg diffuse as
far as 100 μm into the adjacent area, whereas Cu and Zn
only come as far as 30 μm. These values were nearly
doubled (200 μm and 50 μm, respectively) after solution
annealing.
When the values from Table 2 are compared to these
concentration profiles, one can expect a hardness in-
crease in these areas, that have a composition above the

Figure 1 Optical microscopy micrographs of compound cast samples: AlMg1 substrate and Al-alloys with 7% Cu (a, b); 7% Si (c, d); 7% Zn (e, f); and
pure Al (g, h).

Table 2 Lower and upper concentration limits of alloying elements in commercial aluminium alloys
Wrought and casting Al-alloy series mass-% Cu, Si, Zn mass-% Mg
2xxx/2xx 0.1―>7 Cu 0.1―1.8
6xxx/3xx 0.2―>7 Si 0.2―1.5
7xxx/7xx 0.8―>7 Zn 1.5―4.0

48 Konrad J.M. PAPIS et al. Sci China Ser E-Tech Sci | Jan. 2009 | vol. 52 | no. 1 | 46-51

Downloaded to IP: 192.168.0.213 On: 2019-12-10 12:42:46 http://engine.scichina.com/doi/10.1007/s11431-008-0345-9

Figure 2 Concentration profiles after solidification (dashed lines) and after solution annealing (solid lines). The substrate is to the left of the interface; the
droplets are on the right; (a) AlCu7, (b) AlSi7, (c) AlZn7.

lower limits in Table 2. For T6 conditions, diffusion
zone widths of 30 μm (Cu in AlMg1/AlCu7) and 100 μm
(Si in AlMg1/AlSi7) into the substrate, and 100 μm (Mg
in AlMg1/AlCu7) and 70 μm (Mg in AlMg1/AlSi7) into
the droplet can be predicted. For T5, these zones are less
wide, with values of 20, 50, 50 and 30 μm in the same
order as before. Not enough magnesium was provided
from the substrate to form a significant amount of η′
(Zn2Mg) precipitates in AlMg1/AlZn7 couples, which
leads to a small hardness increase only at the interface.
The hardness profile for the AlMg1/AlCu7 couple is
shifted towards the droplet side, and for the AlMg1/
AlSi7 couple slightly towards the substrate. The meas-
ured hardness profiles of all the couples at T5 and T6
states are shown in Figure 3.
At the T6 state (Figure 3(a)), hardness profiles are
shown half a millimeter to either side of the interface.
One would expect the hardness peaks to lie on the inter-
face, but consulting Figures 2 and 3, this is only the case
for AlMg1/AlZn7 couples, where the second phase is
located at the interface only. The optimum ratio of al-
loying elements for the formation of S′ precipitates in
AlMg1/AlCu7 couples lies at the interface, but the peak
is shifted towards the droplet. This can be explained by
the presence of Q′ (Al2Cu) precipitations in the droplet
area. From the composition of β′ (Mg2Si), one would
expect the hardness peak to lie on the substrate’s side of
the interface, which is not the case. This can be ex-
plained by the optimum ratio of magnesium to silicon in
incoherent β″ precipitations of 1:1[15].
A peak in the substrate of the reference couple indi-
cates the partial melt-up of the material with a partially
eutectic solidification, leading to a hardness increase of
about 100 μm into the bulk, and is even more pro-
nounced without solution annealing (Figure 3(b)).
At the T5 state of Al–Al couples, the hardness pro-
files are displayed in a region of 150 μm to both sides of
the interface, the substrate and the droplet material. The

Figure 3 Hardness profiles T6 (a) and T5 (b), measured across the interface (former substrate surface at position 0) of all sample compositions.

Konrad J.M. PAPIS et al. Sci China Ser E-Tech Sci | Jan. 2009 | vol. 52 | no. 1 | 46-51 49

Downloaded to IP: 192.168.0.213 On: 2019-12-10 12:42:46 http://engine.scichina.com/doi/10.1007/s11431-008-0345-9

Cu containing samples are the hardest both in the sub-
strate (due to a higher ageing temperature compared to
the other couples) and in the droplet. This is due to the
formation of AlCu species, as described for T6 condi-
tions. There, the droplet even reached a higher hardness.
A gradual change in hardness from substrate to the
droplet was measured, with the interface hardness lying
in between these values (75 HV), indicating that no sig-
nificant blending of the alloying elements had taken
place (compared to T6, where there is a peak of hardness
at the interface). AlMg1/AlSi7 couples show a little peak
at the interface. This indicates a small amount of inter-
diffusion of Si and Mg, forming MgSi precipitates as
described above. The same is true for Zn containing cou-
ples, which exhibit the peak hardness at the interface.
This big difference at the interface is explained by the
appearance of a second phase after casting. A very dif-
ferent image can be drawn for the samples with pure
aluminium. The substrate was once again partially
melted to a depth of about 150 μm upon contact with the
droplet, causing this commercial alloy to solidify with a
certain amount of eutectic phase. There, the hardness
reaches a high maximum (86 HV) compared to the sub-
strate and droplet values of 50 HV and 21 HV, respec-
tively.
3.2 Al–Mg: Wetting experiments and interface
characterization
Keeping in mind that a ‘compound cast’ structure re-
quires a metallurgically continuous bonding (i.e. free
from cracks or oxides), the manganese coating is essential
for Al–Mg compounds, for various reasons: With its rela-
tively high melting point, its sufficient solubility in Mg
――which ensures a good wettability by that metal――
and without forming IMPs, the Mn coating acts as a good
connection to the melt droplet. Towards the Al- based
substrate, IMPs may form, but because this reaction
happens through diffusion in the solid state, the com-
pound’s integrity is assured, when they are kept thin.
As shown in Figure 4, a layer of IMPs starts to grow
in-between the AlMg1 substrate and the Mn coating at
process temperatures (around 700℃). It becomes obvi-
ous through EDX measurements, that there is a gradient
in Mn and Al concentrations through this ~6 μm thick
IMP layer and in the Mn coating. The Al–Mn phase dia-
gram shows, that multiple IMPs (Al8Mn, λ, μ, Al11Mn4,
γ2) may form. Analysis of the obtained EDX-data re-
50 Konrad J.M. PAPIS et al. Sci China Ser E-Tech Sci | Jan. 2009 | vol. 52 | no. 1 | 46-51
Downloaded to IP: 192.168.0.213 On: 2019-12-10 12:42:46 http://engine.scichina.com/doi/10.1007/s11431-008-0345-9
veals the presence of mainly Al8Mn and γ2 at the borders
to the bulk Al substrate and the Mn layer, respectively.
The μ and Al11Mn4 IMPs appear as well, which are hard
to detect with EDX. The BSE (backscattered electron)
image (Figure 4) suggests the presence of at least three
of these intermetallics. In the β-Mn area, the concentra-
tion of Al ranges from 2 to 16 at-%. The amount of
oxygen in this layer is below 5%.
Figure 4 SEM image of the Al-Mn-Mg interface, featuring a 6-μm thick
IMP layer. On the right, EDX mappings for Mg, Mn and Al are shown.
4 Conclusions
Compound casting of various light metals was investi-
gated on a laboratory scale. Through adequate surface
treatments and coatings, the AlMg1 substrate’s wettabil-
ity was improved in a way that the couples of Al–Al and
Al–Mg were successfully produced. Interfaces showed
very low (Al–Mg) to no (Al–Al) formation of IMPs, and
other defects, such as oxide inclusions, contraction cavi-
ties or cracks.
The combined coating of zincate treatment and elec-
trolytically-deposited zinc for Al–Al compounds offers
major advantages compared to other approaches to join-
ing light metals[1,16,17]. Zones with heat-treatable compo-
sitions were formed in regions ranging up to 200 μm to
both sides of the interface, which could be quantified by
microhardness indents after heat-treatment procedures.
Mechanical properties of the interfacial area can thus be
adapted to the field of a future component′s application.
DICTRA simulations yielded diffusion zone dimensions
of about 50―200 μm, which coincides well with the
experimental findings, and can help to reduce the vol-
ume of future wetting experiments with other alloys.
The zincate treatment is the most important process
step also for surface preparation of the substrate for
Al–Mg compounds. This treatment is followed by elec-
trolytic deposition of an Mn layer of several microns
thickness to protect the substrate from liquefaction by
the Mg melt, without sacrificing too much of wettability.
A thin layer of IMPs forms during couple production,
which might affect mechanical integrity. Keeping this
interface thin is a possible way to improve the com-
pound’s properties.
1 Volder J P. An overview of a number of casting processes. Aluminium
1993, 4: 11―16
2 Pearson T, Wake S J. Improvements in the pretreatment of aluminium
as a substrate for electrodeposition. Trans Inst Met Finish, 1997, 75:
93―97
3 Saito M, Maegawa T, Homma T. Electrochemical analysis of zincate
treatments for Al and Al alloy films. Electrochim Acta, 2005, 51:
1017―1020[DOI]
4 Qi G, Chen X, Shao Z. Influence of bath chemistry on zincate mor-
phology on aluminium bond pad. Thin Solid Films, 2002, 406: 204―
209[DOI]
5 Tang Y C, Davenport A J. The effect of heat treatment and surface
roughness on the zincate treatment of aluminium alloy 6082. Trans
Inst Met Finish, 2001, 79: 85―89
6 Robertson S G, Ritchie I M, Druskovich D M. A kinetic and electro-
chemical study of the zincate immersion process for aluminium. J
Appl Electrochem, 1995, 25: 659―666
7 Monteiro F J, Barbosa M A, Ross D H, et al. Pretreatments to im-
prove the electrodeposits on aluminium. Surf Interface Anal, 1991,
17: 519―528[DOI]
Konrad J.M. PAPIS et al. Sci China Ser E-Tech Sci | Jan. 2009 | vol. 52 | no. 1 | 46-51
Downloaded to IP: 192.168.0.213 On: 2019-12-10 12:42:46 http://engine.scichina.com/doi/10.1007/s11431-008-0345-9
8 Zipperian D C, Raghavan S, Pritzker M D. Physical and chemical
characteristics of the zincate immersion process for aluminium and
aluminium alloys. J Electrochem Soc, 1987, 8B: C437―C438
9 Chen P-Y, Hussey C L. The electrodeposition of Mn and Zn-Mn al-
loys from the room-temperature tri-1-butylmethylammonium bis
((trifluoromethane)sulfonyl)imide ionic liquid. Electrochim Acta,
2007, 5: 1857―1864[DOI]
10 Davis J R, ed. ASM Speciality Handbook―Aluminium and Alumin-
ium Alloys. Materials Park, OH: ASM International, Materials Park,
OH 440730002, USA, 1993
11 Andersson J O, Helander T, Hoglund L, et al. Thermo-Calc &
DICTRA, computational tools for materials science. Calphad, 2002,
26: 273―312[DOI]
12 Ansara I, Dinsdale A T, Rand MH, eds. Thermochemical Database
for Light Metal Alloys. COST 507 Final Report Round 2, Vol. 2,
Luxembourg, 1998
13 Prikhodovsky A. PhD Thesis, Aachen: RWTH Aachen University,
2000
14 Polmear I J. Light Alloys――Metallurgy of the Light Metals. Hon-
eycombe R, Hancock P, eds. London: Arnold, 1995
15 Murayama M, Hono K. Pre-precipitate clusters and precipitation proc-
esses in Al-Mg-Si alloys. Acta Mater, 1999, 47: 1537―1548[DOI]
16 Ronnhult T, Rilby U, Olefjord I. The surface state and weldability of
aluminium alloys. Mater Sci Eng, 1980, 42: 329―336[DOI]
17 Borrisutthekul R, Miyashita Y, Mutoh Y. Dissimilar material laser
welding between magnesium alloy AZ31B and aluminium alloy
A5052-O. Sci Techn Adv Mater, 2005, 6: 199―204[DOI]
51