Chlorination of aluminium-copper alloys

Article abstract of DOI:10.1016/j.tca.2007.02.012

The chlorination of Al-Cu alloys is investigated by thermogravimetry, X-ray diffraction, scanning electron microscopy, energy dispersive spectroscopy, and energy dispersive X-ray fluorescence, between 200 and 500 °C. The starting temperature for the chlorination of selected alloys is determined. Results indicate that the reactivity of the alloys is different from that of the pure metals. Analysis of the reaction products reveals that there are interactions between the different chlorides formed. These interactions lead to an enhancement in the copper chlorides volatilization and the formation of well-developed crystals of CuCl₂ instead of the mixture of CuCl₂ and CuCl that is obtained in the chlorination of pure copper. A schematic diagram that describes the chlorination mechanisms for the pure metals and AlCu alloys is presented.

Full text of DOI:10.1016/j.tca.2007.02.012

Thermochimica Acta 457 (2007) 83–91
Chlorination of aluminium–copper alloys
G. De Miccoa,b,∗, D.M. Pasquevich , A.E. Bohe
a,c
c,d
´
a Comisio´n Nacional de Energ´ıa Ato´mica (C.N.E.A.), Avenida Bustillo 9500, 8400 San Carlos de Bariloche, Argentina
b Universidad Nacional de Cuyo, Instituto Balseiro, Avenida Bustillo 9500, 8400 San Carlos de Bariloche, Argentina
c Consejo Nacional de Investigaciones Cient´ıficas y Te´cnicas (CONICET), Argentina
d Universidad Nacional del Comahue, Centro Regional Universitario Bariloche, 8400 San Carlos de Bariloche, Argentina
Received 22 December 2006; received in revised form 6 February 2007; accepted 16 February 2007
Available online 27 February 2007
Abstract
The chlorination of Al–Cu alloys is investigated by thermogravimetry, X-ray diffraction, scanning electron microscopy, energy dispersive
spectroscopy, and energy dispersive X-ray fluorescence, between 200 and 500 ◦C. The starting temperature for the chlorination of selected alloys
is determined. Results indicate that the reactivity of the alloys is different from that of the pure metals. Analysis of the reaction products reveals
that there are interactions between the different chlorides formed. These interactions lead to an enhancement in the copper chlorides volatilization
and the formation of well-developed crystals of CuCl2 instead of the mixture of CuCl2 and CuCl that is obtained in the chlorination of pure copper.
A schematic diagram that describes the chlorination mechanisms for the pure metals and AlCu alloys is presented.
© 2007 Elsevier B.V. All rights reserved.
Keywords: Chlorination; Aluminium–copper alloys; Thermogravimetry
1. Introduction
To the authors’ best knowledge, the reaction of chlorine with
Al–Cu alloys has not been studied. It is only known that the reac-
tion of chlorine with aluminium yields AlCl3 [1–5,13], while
that of chlorine with copper has two possible products, CuCl and
CuCl2 [14]. For copper, in addition to the information related
with synthesis of copper(I) chloride by dry chlorination [7,8,14],
a few basic studies are available [15–18]. Recently, De Micco et
al. [18] reported the formation of both chlorides during the reac-
tion of copper with chlorine for temperatures between 150 and
750 ◦C. For chlorinations up to 250 ◦C they obtserved that cop-
per reacts to form CuCl, which forms CuCl2 by further reaction,
leading to a layered morphology.
Interactions between reaction products and solid reactants
during the chlorination of multi-component systems are known.
[19,20] For that reason, similar behaviours are to be expected
during the chlorination of alloys.
A gaseous complex between aluminium and copper chlo-
rides, namely CuAl2Cl8, was obtained during a process of
etching an aluminium–copper alloy. Its occurrence allows to
convert copper chlorides to a form having sufficient volatility
for removal by complexing with the aluminium chloride. The
complex is readily detected in the gas phase at 280 ◦C [21]. A
ꢀG◦ valueof33.71 kJ/molhasbeenreportedfortheformationof
the gaseous complex from solid CuCl2 and gaseous Al2Cl6 [22].
The reactions of chlorine with pure metals have not been thor-
oughly studied. The majority of the studies are concerned with
production of anhydrous chlorides for industrial applications
[1–7]. Even less is the information available on alloy chlorina-
tion. However, in the last few years, an increasing number of
chlorination studies related with the recovery of valuable metals
suchasCo, Ni, V, W, Mo, Cr, Fe, Cu, andZnfromcomplexscraps
have been published [8–10]. This indicates that the chlorination
process has become an important technological subject.
As far as basic research is concerned, Landsberg studied
the chlorination rates of Si–Ge, Mo–Re and Au–Pt alloys with
respect to temperature and chlorine pressure. He found that the
reactivities of the alloys are different from those of the pure
constituents [11]. Chang and Wei [12] reviewed the corrosion
of several metals and alloys in atmospheres containing chlorine
above 200 ◦C. Regarding kinetics and mechanism he states that
high temperature corrosion consists mainly of diffusion of ions
in the scale of chlorides formed and volatilization of the scale.
∗
Corresponding author at: Centro Ato´mico Bariloche, Bustillo 9500,
R84002AGP, S.C. de, Bariloche, R´ıo Negro, Argentina.
E-mail address: demiccog@cab.cnea.gov.ar (G. De Micco).
0040-6031/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.tca.2007.02.012

84
G. De Micco et al. / Thermochimica Acta 457 (2007) 83–91
Finally, the Raman spectrum of the complex was measured by
Emmenegger et al. for temperatures up to 300 ◦C. They found
that on increasing temperature, the spectrum of CuAl2Cl8(g)
appears with increasing intensity [23].
3. Results and discussion
3.1. Reactivity of Al–Cu alloys
In the present work, the chlorination of Al–Cu alloys was
investigated for the first time. Following the reaction process
it is possible to establish the general stages of the chlorination
and to reveal the interactions that occur between the different
compounds. The effect of temperature and alloy composition
on chloride volatilization is analyzed.
The chlorination of the Al–Cu system was studied by focus-
ing on the chlorination of three alloys. From the thermodynamic
point of view, the negative ꢀG◦ values for the chlorination of
copper and aluminium [26] indicate that these reactions are ther-
modynamically feasible at all temperatures. However, it is well
known that when a naked metal or alloy surface is put in con-
tact with air, an oxide scale is formed, which depending on
its microstructural characteristics will have different protective
properties towards the corrosive agents. This is why it can be
asserted that some metals will not react when they are put in
direct contact with chlorine, until the conditions are appropriate
for its penetration through the passivated oxide film and the con-
tact of the gas with the underlying metal is made. Chlorine can
reach the bulk metal by chlorination of the protective oxide scale
or by diffusion through grain boundaries or through structural
vacancies of the oxide film.
2. Experimental details
2.1. Alloys
Commercial metals (99.9% purity) were used to prepare
the following alloys: Al 46 wt.%–Cu, Al 18 wt.%–Cu and Cu
4 wt.%–Al. A powder of each alloy was prepared by mechanical
abrasion with an analytical mill (IKA Labortechnik). Thermal
treatments at 450 ◦C for 24 and 52 h in air were performed
to release possible residual tensions. The powder samples
of each alloy were well characterized by energy dispersive
X-ray fluorescence (ED-XRF), scanning electron microscopy
(SEM) and X-ray diffraction (XRD). The results are shown in
Fig. 1.
As the figure shows, Al 46 wt.%–Cu and Al 18 wt.%–Cu are
isotropic particles with faceted faces of different sizes from var-
ious microns to about 100 ␮m. Cu 4 wt.%–Al powder consists
of needle-shped particles of about 1000 ␮m long. XRD patterns
obtained show that the structures of the alloys are the intermetal-
lic compound Al2Cu for the Al 46 wt.%–Cu alloy, the gamma
phase (Al4Cu9) for the Al 18 wt.%–Cu alloy, and a mixture of
Al and Al2Cu for the Cu 4 wt.%–Al alloy.
The ꢀG◦ values for the chlorination reactions are negative for
copper oxide, and positive for aluminium oxide, at temperatures
below 1000 ◦C [26]. According to this, copper oxide will react
while aluminium oxide will not. Therefore, in the case of copper,
chlorine can reach the metal either by chlorination of the oxide
scale or by diffusion through the scale depending on the relative
velocities of the two mechanisms. On the other hand, aluminium
will be attacked by chlorine only when diffusion of chlorine
through the oxide scale takes place.
3.1.1. Non-isothermal chlorination of Al, Cu, Al
46 wt.%–Cu, Al 18 wt.%–Cu, and Cu 4 wt.%–Al
The initial reaction temperature which is associated with the
system reactivity was determined by non-isothermal thermo-
gravimetric measurements. The chlorination curves for each
alloy together with the chlorination of the pure metals are
shown in Fig. 2. This figure shows the ratio between the mass
change and the initial mass of the sample as a function of
temperature.
InFig. 2, wecanseethatthestartingtemperaturesforthereac-
tions of the copper-rich and aluminium-rich alloys are similar to
thoseofthecorrespondingpuremetals, whiletheAl46 wt.%–Cu
alloy starts at a higher temperature.
2.2. Chlorination
The gases used were Cl2 99.8% purity (Indupa, Argentina)
and Ar 99.99% purity (AGA, Argentina). Isothermal and
non-isothermal chlorination reactions were carried out in a ther-
mogravimetric analyzer (TGA), which is extensively described
elsewhere [24]. This thermogravimetric analyzer consists of
an electrobalance (Model 2000, Cahn Instruments, Inc.) suit-
able for working with corrosive atmospheres, a gas line and
an acquisition system. This experimental set-up has a sensitiv-
ity of 5 ␮g while operating at 950 ◦C under a flow of 8 l/h.
Samples of 20 mg were placed in a quartz crucible inside the
reactor in an argon flow of 1.3 l/h. To begin the non-isothermal
reactions, a chlorine flow of 0.8 l/h was introduced and, at the
same time, the heating started with a ramp of 100 ◦C/h. For the
isothermal reactions, the solids were heated until they reached
the desirable temperature before chlorine injection. The partial
pressure of chlorine in the Ar–Cl2 mixture was 35.5 kPa. Over
the experimental conditions described above, the rate order of
the alloy chlorinations is 10−6 mol of Cl2/s, which reveals a
gaseous phase diffusion control as demonstrated by previous
publications [18,25]. Therefore, the influence of particle size in
the reaction rate was considered negligible.
For the Al 18 wt.%–Cu alloy, the reaction starts at about
100 ◦C, the same for copper. The chlorination of Cu 4 wt.%–Al
starts at about 175 ◦C, and this temperature is 45 ◦C higher than
the starting temperature for pure aluminium (130 ◦C). Besides,
the mass changes observed in the chlorination of these alloys are
in accordance with the mass changes observed in the chlorina-
tion of the constituents. The copper-rich alloy presents a mass
gain and the same happens in the chlorination of copper due to
the formation of condensed copper chlorides. The chlorination
of the aluminium rich alloy shows a mass loss, also observed
in the chlorination of aluminium due to the volatilization of
aluminium chlorides.
The chlorination of the Al 46 wt.%–Cu alloy starts at 230 ◦C,
while reactions of the pure metals start at 100 ◦C for copper and

G. De Micco et al. / Thermochimica Acta 457 (2007) 83–91
85
Fig. 1. Morphology, composition and structure of the initial samples: (a) and (d) correspond to powder of Al 46 wt.%–Cu alloy, (b) and (e) powder of Al 18 wt.%–Cu
alloy, and (c) and (f) powder of Cu 4 wt.%–Al alloy. XRD reference patterns from Ref. [28].
at 130 ◦C for aluminium, as mentioned before. Moreover, the
mass change observed in the alloy chlorination is very different
from that of the pure metals. To investigate whether the constant
mass observed in the first part of the curve, i.e. below 230 ◦C, is
due to the absence of reaction or is the result of the compensation
between condensed copper chlorides and gaseous aluminium
chlorides, we did an isothermal chlorination at 200 ◦C for 2 h.
In that experiment, no mass change was observed. Photographs
of the initial and chlorinated samples obtained with SEM are also
shown in Fig. 2a and b, respectively, and they reveal that there
were no changes in the sample morphology. Finally, XRD anal-
ysis of the same sample confirmed that there was only Al2Cu.
This behaviour is different from that observed in the chlorina-
tion of the pure metals because both begin to react at 200 ◦C.
The reactivity of the alloy at this temperature is thus lower than
that of the pure metals.
3.2. Chlorination of Al 46 wt.%–Cu (Al2Cu)
To study the general way in which the chlorination reac-
tions of Al–Cu alloys take place, we analyzed in particular the
chlorination of Al2Cu.

86
G. De Micco et al. / Thermochimica Acta 457 (2007) 83–91
Fig. 3. Isothermal TG curve for the chlorination of Al 46 wt.%–Cu at 400 ◦C.
(a) Reaction products, (b) reaction quenched during the mass loss, indicated by
ꢀ, (c) and (d) reaction quenched in the minimum, indicated by ᭹.
Fig. 2. Non-isothermalTGcurvesforthechlorinationofCu, Al, Al46 wt.%–Cu,
Al 18 wt.%–Cu, and Cu 4 wt.%–Al. (a) Initial sample for the chlorination of Al
46 wt.%–Cu, (b) unreacted Al 46 wt.%–Cu alloy after 2 h of chlorination at
200 ◦C.
followed by a mass gain, after which, the mass remains constant.
We analyzed the chlorination residues that stayed in the crucible
at the end of the reaction by different techniques. XRD patterns
(Fig. 4a) showed the presence of only CuCl2, ED-XRF and EDS
analyses confirmed that the amounts of copper and chlorine cor-
respond to CuCl2, and a small amount of aluminium was also
detected (between 0.5 and 2 wt.%). SEM examinations (Fig. 3a)
showed a uniform morphology of needles already observed for
CuCl2 [27].
According to these results, during the chlorination reaction
almost all aluminium forms volatile chlorides that evaporate.
This can be explained since the vapour pressure of AlCl3 is
10−2 kPa at 75 ◦C (see Table 1). However, there is also part
of the initial copper that evaporates, as will be discussed sub-
sequently. This is an unexpected result, taking into account
that the vapour pressure of copper chlorides reaches a value
of 10−2 kPa at temperatures considerably higher, as shown
in Table 1.
As was reviewed by Chang and Wei [12], high temperature
chlorine corrosion rate has two contributions: diffusion of ions
throughthechloridescale, whenalayeredproductisformed, and
volatilization of the scale. Therefore, typical curves of weight
change vs. time for metals show a parabolic mass gain dur-
ing the growth period of metal chloride if volatilization effect
is not significant, followed by a linear mass loss when growth
of chloride is not fast enough to supply the amount required
for volatilization. Usually, diffusion of ions through the chlo-
ride scale is fast enough only at temperatures above 0.5 Tm,
(Tm: melting point), and volatilization is significant only at
vapour pressures above 10−2 kPa (10−4 atm). The values of
Tm and T4 (the temperature at which the vapour pressure is
10−4 atm) for copper and aluminium chlorides are presented
in Table 1.
In order to understand how the alloy chlorination occurs,
we proceeded to interrupt the chlorination reactions at different
During copper chlorination reactions, formation of a product
layer consisting of CuCl and CuCl2 has been reported [18]. For
temperatures below 275 ◦C (about 0.5 Tm for CuCl2) this layer
tends to stop the reaction by blocking ion diffusion. On the other
hand, in the chlorination of aluminium, volatilization of AlCl3
does not allow the formation of a protective scale.
The TG curve for the chlorination of Al2Cu at 400 ◦C is
shown in Fig. 3. The shape of this curve is entirely different
from the typical mass changes versus time curves described
before, which put in evidence that there are several simultane-
ous processes taking place. The reaction starts with a mass loss
Table 1
Aluminium and copper chlorides melting point (Tm) and temperatures (T4) at
which the chloride vapour pressure is 10−2 kPa (10−4 atm)
Chloride
Tm (◦C)
T4 (◦C)
Reference
AlCl3
CuCl
CuCl2
181 sublimates
422
500
75
950
Decomposes
[13,25]
[14,25]
[18]
Fig. 4. XRD patterns obtained at different stages during the chlorination of Al
46 wt.%–Cu at 400 ◦C. XRD reference patterns from Ref. [28].

G. De Micco et al. / Thermochimica Acta 457 (2007) 83–91
87
Fig. 5. Schematic diagram for the chlorination evolution of copper, aluminium and Al–Cu alloys.
reaction degrees (as shown by symbols in Fig. 3) and to ana-
3.3. Influence of temperature in the chlorination of Al–Cu
alloys
lyze the samples by XRD and SEM. SEM photographs obtained
are shown in Fig. 3(b–d), and XRD patterns are presented in
Fig. 4. In the first part of the curve during the mass loss, XRD
patterns (Fig. 4b) show that there is Al2Cu, Cu and small peaks
that may correspond to AlCl3·6H2O. The presence of remain-
ing AlCl3 can be understood considering that the sample was
quenched before the chlorination is completed. Due to the fast
temperature decrease, part of the AlCl3 which is being formed,
is not able to volatilize and remains with the solid product. The
occurrence of metallic copper means that aluminium is being
eliminated faster than copper by volatilization of its chloride.
However, copper chlorination does occur as well during the first
part of the reactions, the presence of CuCl2 was confirmed by
interrupting and quenching several partially chlorinated samples
at different temperatures. SEM micrographs (Fig. 3b) also reveal
that the reaction proceeds through the formation of micro cracks
in the solid alloy, and the aluminium chloride formed escapes
from those cracks.
We did isothermal chlorinations of the different alloys in the
range between 175 and 500 ◦C, and chlorine partial pressure
of 35.5 kPa. The influence of temperature on the chlorination
reactions can be observed in Fig. 6 where the TG curves are
presented.
The aluminium rich alloys (Fig. 6a) show a mass loss during
the whole reaction which corresponds to volatilization of the
chlorides formed. As Fig. 6a also shows, there is an initial time
interval in which the mass loss rate is null or very slow. This
time gap is smaller as temperature increases. Similar delays at
the beginning of the reaction were observed in aluminium chlo-
rination, especially for temperatures below 250 ◦C. The initial
At the minimum of the curve, the morphology of the sample
is completely different (Fig. 3c) there is a sintering pow-
der of metallic copper with a small content of aluminium
(about 12 wt.%, according to EDS measurements), and there are
also some CuCl2 crystals. XRD profile shows mainly copper
(Fig. 4c).
A curious feature of this curve is the fluctuations observed
between 400 and 600 s. This behaviour may be due to the forma-
tion of condensed chlorides (mainly from copper) that stop the
reaction by preventing chlorine to contact the remaining metal,
at the point where the rate of mass gain reaches a zero value, the
volatilization rate becomes clear. After some seconds of mass
loss, the chloride layer thins and allows chlorine to reach fresh
metal, leading again to a subsequent mass gain.
Fig. 6. Isothermal TG curves for the chlorination of (a) Cu 4 wt.%–Al between
175 and 400 ◦C, (b) Al 46 wt.%–Cu between 250 and 500 ◦C and (c) Al
18 wt.%–Cu between 250 and 400 ◦C.
A schematic diagram for the reaction of copper, aluminium
and Al–Cu alloys is presented in Fig. 5.

88
G. De Micco et al. / Thermochimica Acta 457 (2007) 83–91
time interval may correspond to an induction time due to the
diffusion of chlorine through the oxide scale.
The total mass loss corresponds to about 98% of the initial
mass of the sample. The difference is due to the non-reactive
aluminium oxide scale and also to some copper chlorides as
will be discussed in the next section.
SEM analyses of the chlorinated samples allowed us to inves-
tigate the morphologies of the reaction products. Fig. 7a–e
correspond to the samples after chlorination of copper at 300 ◦C,
Al 18 wt.%–Cu at 250 ◦C, Al 46 wt.%–Cu at 400 ◦C, mix-
ture of Al and Cu powders in the ratio of Al2Cu at 400 ◦C,
and Cu 4 wt.%–Al at 250 ◦C, respectively. Only one temper-
ature is presented for each system because the morphologies
of the final samples were essentially the same in the tempera-
ture range studied. As mentioned above, XRD of these samples
indicated that they consist of CuCl2 in the case of the alloy
chlorination, whereas for copper chlorination CuCl was also
detected.
SEM observations revealed that CuCl2 grows in needles and,
the morphology of the chlorides obtained in the chlorination
of alloys, as well as in pure copper, is the same. However, the
CuCl2 crystal growth is more important during the chlorination
of Al–Cu systems (Fig. 7b–e) than in the chlorination of pure
copper (Fig. 7a).
The TG curves for the chlorination of Al 46 wt.%–Cu alloys
are shown in Fig. 6b. Again, for temperatures below 400 ◦C,
the reaction does not start immediately, as evidenced by an
initial time span with no-mass-change. This behaviour is simi-
lar to that of aluminium, as mentioned before, and is different
from that observed in the chlorination of copper powder for the
same conditions of gaseous flow and partial pressure of chlo-
rine. In that case, the mass gain starts immediately after chlorine
reaches the metal for temperatures above 200 ◦C. Another char-
acteristic of the TG curves that can be observed in Fig. 6b is
that, as temperature increases, the initial mass loss is larger
and the amount of the final sample is smaller. For temperatures
above 500 ◦C, at the end of the curve the mass continuously
decreases until complete volatilization, whereas for other tem-
peratures the mass reaches an almost constant value. This is in
agreement with the volatilization detected in copper chlorination
reactions where a similar behaviour was observed [18]. Finally,
in Fig. 6c the chlorination TG curves for the Al 18 wt.%–Cu
alloy at three temperatures are presented. These reactions show a
mass gain typical of condensed copper chloride formation. How-
ever, simultaneous evaporation of volatile chlorides does occur
as evidenced by the low mass gain observed, and confirmed by
analysis of the final condensed products, as will be discussed
subsequently.
The atomic percentage of aluminium contained in the copper
chlorides was determined by ED-XRF and EDS for the different
alloys, and the results are shown in Table 2. It can be seen that
the values are below 5% and they seem not to be influenced by
temperature or alloy composition. The amounts of aluminium
contained in the final samples are attributed to the non-reactive
aluminium oxide scale.
3.4.1. Enhanced volatilization of copper chlorides
It is possible to calculate the amount of chlorides evaporated
during the chlorination reactions from the difference between
the initial and final masses of each metal. To do that, we mea-
sured the final composition corresponding to the chlorination of
Al 46 wt.%–Cu and Al 18 wt.%–Cu, and a mixture of Al and Cu
powders in the ratio of Al2Cu, at different temperatures with ED-
XRF (the corresponding results for one temperature are shown
also in Fig. 7f). The amount of the final samples after chlorina-
tion of Cu 4 wt.%–Al were not large enough to be measured by
ED-XRF and, consequently, their compositions were analyzed
by EDS.
In Table 3, the percentages of the initial mass of copper
volatilized during the chlorination of Al–Cu systems compared
with metallic copper [18] are shown. Taking into account that the
volatilization of copper chlorides is negligible during the chlori-
nation of pure copper (see Table 3), our results clearly show that
the presence of aluminium enhances the volatilization of copper
chlorides during the chlorination reactions. This enhancement
can be understood by considering the formation of a gaseous
complex between Al2Cl6 and CuCl2, according to the following
3.4. Reaction products from the chlorination of Cu and
Al–Cu alloys
The chlorinated samples from the reaction of Al 46 wt.%–Cu
and Al 18 wt.%–Cu were analyzed by XRD. We found that they
consist of CuCl2 for all temperatures, except for 500 ◦C (in the
case of Al 46 wt.%–Cu) where CuCl was also detected. To avoid
CuCl2 decomposition after the reaction at the highest temper-
atures, i.e. at 450 and 500 ◦C, the samples were cooled in Cl2
atmosphere [18].
The absence of CuCl at the lower temperatures is in accor-
dance with thermodynamical data [26]. The ꢀG◦ values for
the reaction 2CuCl + Cl2(g) → 2CuCl2 are negative for tempera-
tures up to about 400 ◦C where it has a value of −3.295 kJ/mol of
Cl2, indicating that the most stable specie is CuCl2 below that
temperature. However, in the reaction of pure copper powder
with chlorine at 200 and 300 ◦C, we found that the chlori-
nated sample consisted of a mixture of CuCl and CuCl2. The
presence of CuCl in that case was attributed to a deficit of
chlorine established by the formation of solid CuCl2, which
retards or prevents the contact between Cl2 and CuCl [18]. In
the alloy chlorination, a possible explanation for the absence
of CuCl is that the continuous evaporation of AlCl3 from the
solid does not allow the formation of that protective layer of
CuCl2 and gives a path for chlorine to reach CuCl and form
CuCl2.
Table 2
Percentage of aluminium in the final samples corresponding to the chlorination
of different alloys at temperatures between 250 and 400 ◦C
Temperature (◦C) Percentage of aluminium in the final copper chlorides
Al 18 wt.%–Cu Al 46 wt.%–Cu Cu 4 wt.%–Al
250
300
400
1.6
2
2
2
1.4
1
4.5
5

G. De Micco et al. / Thermochimica Acta 457 (2007) 83–91
89
Fig. 7. Final samples from the chlorination of: (a) copper at 300 ◦C, (b) Al 18 wt.%–Cu at 250 ◦C, (c) Al 46 wt.%–Cu at 400 ◦C, (d) mixture of Al and Cu powders
in the ratio of Al2Cu at 400 ◦C, (e) Cu 4 wt.%–Al at 250 ◦C and (f) atomic composition of the samples determined by EDS and ED-XRF.
reaction [23]:
probably being formed from an isotropic media through a
volatilization–crystallization mechanism during the reactions.
With a simple calculation based on the moles of copper
lost during each reaction, and considering that all copper is
volatilized as CuAl2Cl8, it is possible to estimate the average
partial pressure of the complex under the experimental con-
ditions. For example, in the chlorination of Al 46 wt.%–Cu at
CuCl2(s) + Al2Cl6(g) = CuAl2Cl8(g)
(1)
The presence of a high concentration of copper in the gaseous
phase as CuAl2Cl8, can explain the well defined crystals of
CuCl2 obtained, which are bigger than those observed in
the chlorination of copper (see Fig. 7). These chlorides are
Table 3
Percentage of copper volatilized for different metals and alloys at different temperatures, and the corresponding quantity: SCu/2Al
T (◦C)
Percentage of the initial mass of copper volatilized in
the chlorination of different metals an alloys
Moles of Cu that were volatilized every two moles of Al
volatilized (SCu/2Al
)
Cua
Al 18 wt.%–Cu
Al 46 wt.%–Cu
Cu 4 wt.%–Al
Al 18 wt.%–Cu
Al 46 wt.%–Cu
Cu 4 wt.%–Al
250
300
400
Non-detected
Non-detected
<0.5
23
22
26
19
22
32
29b
84
100
95
∼1
0.19
0.22
0.32
0.30b
0.03
0.03
0.03
0.98
∼1
a
Ref. [18].
b
Corresponds to the chlorination of a mixture of powders of copper and aluminium in the ratio of Al2Cu.

90
G. De Micco et al. / Thermochimica Acta 457 (2007) 83–91
400 ◦C, P(CuAl2Cl8) is about 10−1 kPa (10−3 atm). The results
obtained are in good agreement with the values of P(CuAl2Cl8)
calculated taking into account the ꢀG◦ values for reaction (1)
[22], P(Al2Cl6) 102 kPa (1 atm), and the same experimental con-
ditions. Although the chlorinating gas is a mixture of Cl2–Ar,
in the sample surroundings AlCl3 is being generated and Cl2 is
being exhausted by the reaction, for that reason the partial pres-
sure of AlCl3 is expected to be high. Moreover, the residence
time of AlCl3 in the sample is long due to the low gaseous flow
(2.1 l/h) and because the AlCl3 formed has to diffuse through a
CuCl2 layer before reaching the gaseous phase, as schematically
shown in Fig. 5.
this alloy even a low partial pressure of the complex represents
a high portion of the initial mass of copper.
Regarding the effect of temperature, the partial pressure of the
complex increases with temperature, as can be seen in Table 3
were there is a slight increase in the percentage of copper chlo-
rides being volatilized with temperature. This is to be expected
since the reaction given by Eq. (1) is endothermic [23]. Further-
more, as is also shown in Table 3, the volatilization is the same
for the case of the Al 46 wt.%–Cu alloy chlorination and for
the mixed powders of pure metals. This indicates that the only
interactions which are important in the complex formation are
those between gaseous aluminium chloride and the surface of
solid CuCl2.
To compare the behaviour of the different Al–Cu systems, it
is useful to calculate the moles of copper that are volatilized for
every 2 mol of aluminium during each chlorination reaction:
4. Conclusions
(moles of Cu volatilized during the chlorination)
The initial temperature for the reaction of selected Al–Cu
alloys with chlorine was experimentally determined. For the Al
18 wt.%–Cu and Cu 4 wt.%–Al alloys, the behaviour observed
was similar to that of pure copper and aluminium, respectively.
On the other hand, for the Al 46 wt.%–Cu alloy, a significant
increase in the reaction temperature was found, which can be
due to a change in the microstructure of the oxide scale.
The chlorination of Al 46 wt.%–Cu at 400 ◦C was analyzed
in detail. We found that CuCl2 is the only solid product. At
intermediate stages of the reaction, we detected metallic copper
which indicates that aluminium is being removed from the alloy
by the volatilization of its chloride faster than copper.
The needle morphology of the CuCl2 obtained in the
chlorination of alloys is the same that was observed in
copper chlorination reactions. However, in the case of the
alloys, the crystals are much bigger. This was attributed to
the high partial pressure of CuAl2Cl8 which would lead to
a volatilization–crystallization growth mechanism during the
reaction.
sCu/2Al
=
2 × (moles of Al volatilized during the chlorination)
This quantity is called SCu/2Al and is also shown in Table 3.
Therefore, taking into account Eq. (1), SCu/2Al ∼ 1 means that
all aluminium that left the sample has formed CuAl2Cl8. On the
other hand, low values of SCu/2Al (SCu/2Al ∼ 0) indicate that most
of the aluminium left the sample as Al2Cl6 without forming the
gaseous complex.
When moving from left to right in Table 3, due to an increase
in the relative content of aluminium in the alloys, an increase
in the partial pressure of Al2Cl6 is expected. Meanwhile, owing
to the lower percentages of copper in the alloys which lead to
the formation of less CuCl2, there would be a decrease in the
surface of reaction for the formation of the gaseous complex.
Therefore, the SCu/2Al values obtained for each alloy may give
evidence on which of the above mentioned opposite behaviours
is determining the amounts of complex formed. In the Al
18 wt.%–Cu alloy, the value of SCu/2Al ∼ 1 indicates that all the
aluminium chloride available is forming CuAl2Cl8 and this is
expected since Al2Cl6 is the limiting reactant considering the
reaction given by Eq. (1) and the alloy composition. On the other
hand, for the Al 46 wt.%–Cu alloy, which has the same Cu/Al
ratio of the complex, the low values of SCu/2Al give evidence that,
in this case, CuCl2 is the limiting specie in the complex forma-
tion reaction. According to these results, the maximum partial
pressure of AlCu2Cl8 would occur at some intermediate alloy
composition between Al 46 wt.%–Cu and Al 18 wt.%–Cu. The
aluminium content of alloy has to be as high as possible to pro-
duce the maximum Al2Cl6 partial pressure in combination with
a high reaction surface, i.e. enough amount of CuCl2 available
for the reaction.
Finally, regarding chloride interactions, we found that
volatilization of copper chlorides is enhanced during the alloy
chlorination. The volatilization observed for the different alloys
can be understood taking into account the CuAl2Cl8 forma-
tion reaction which involves solid CuCl2 reacting with gaseous
Al2Cl6.
Acknowledgements
The authors would like to thanks the Agencia Nacional
´
´
´
de Promocion Cientıfica y Tecnologica (ANPCyT), Consejo
´
´
Nacional de Investigaciones Cientıficas y Tecnicas (CONICET)
and Universidad Nacional del Comahue for the financial support
of this work.
Finally, for the Cu 4 wt.%–Al alloy, the initial amount of
copper is below the stoichiometric quantities of the complex,
copper being the limiting reactant in this case. The surface of
interaction between CuCl2 and Al2Cl6 is significantly reduced,
leading to a very low partial pressure of CuAl2Cl8, as can be seen
from the values of SCu/2Al ∼ 0. The high percentage of initial
mass of copper volatilized for this alloy are not related with
a great amount of complex formed, but are due to the initial
composition of alloy which has a small amount of copper. For
References
[1] R.L. De Beauchamp, Preparation of Anhydrous Aluminium Chloride, Inf.
Circ. U.S. Bur. Mines, no 8412, 1969.
[2] C. Furnas (Ed.), Roger‘s Industrial Chemistry, vol. 1, Van Nostrand, New
York, 1942, p. 460 (ch. 2).
[3] Synthesis of Aluminium Chloride. Report from the internet:
http://www.rhodium.ws.

G. De Micco et al. / Thermochimica Acta 457 (2007) 83–91
91
[4] L.S. Belknap, Aluminum trichloride production, US Pat. 3,721,731, Cabot
[15] R. Tit´ı-Manyaka, I. Iwasaki, Trans. SME/AIME 260 (1976) 282–288.
[16] W. Sesselmann, T.J. Chuang, Surf. Sci. 176 (1986) 32–66.
[17] W. Sesselmann, T.J. Chuang, Surf. Sci. 176 (1986) 67–90.
[18] G. De Micco, A.E. Bohe´, D.M. Pasquevich, J. Alloys Compd., in press
doi:10.1016/j.jallcom.2006.08.
Corporation, Boston, Massachusetts, USA, 1973.
[5] L. Piccolo, M. Ghirga, B. Calcagno, Process for preparing aluminium
trichloride, US Pat. 3,812,241, Societa Italiana Resine S.I.R.S.P.A., Milan,
Italy, 1974.
[6] J.P. Remeika, B. Batlogg, Mater. Res. Bull. 15 (1980) 1179–1182.
[7] N. Bourhila, N. Thomas, J. Palleau, J. Torres, C. Bernard, R. Madar, Appl.
Surf. Sci. 91 (1995) 175–181.
[8] I. Gaballah, M. Djona, J.C. Mugica, R. Solozabal, Resour. Conserv. Recy-
cling 10 (1–2) (1994) 87–96.
[9] R.E. Siemens, B.W. Jong, J.H. Russell, Conserv. Recycling 9 (2) (1986)
189–196.
[10] K. Matsumaru, M. Susa, K. Nagata, J. Iron Steel Inst. Jpn. 82 (10) (1996)
1–6.
[19] J.A. Gonza´lez, F.C. Gennari, M. delCarmenRuiz, A.E. Bohe´, D.M. Pasque-
vich, Trans. Inst. Min. Met. Section C 107 (1998) 130–138.
[20] A.E. Bohe´, J.J. Andrade Gamboa, D.M. Pasquevich, Mater. Sci. Technol.
13 (1997) 865–871.
[21] N.J. Norvell, Process for etching an aluminium–copper alloy, U.S. Patent
4,468,284, http://www.freepatentsonline.com/4468284.html.
[22] A. Dell’Anna, F.P. Emmenegger, Helv. Chim. Acta 58 (1975) 1145.
[23] F.P. Emmenegger, C. Rohrbasser, W. Schla¨pfer, Inorg. Nucl. Chem. Lett.
12 (1976) 127–131.
[11] A. Landsberg, J. Less-Common Met. 159 (1990) 163–172.
[12] Y.N. Chang, F.I. Wei, J. Mater. Sci. 26 (1991) 3693–3698.
[13] O. Helmboldt, L. Keith Hudson, C. Misra, K. Wefers, H. Stark, M.
Danner, Ullman’s Encyclopedia of Industrial Chemistry, sixth ed., Elec-
tronic Release. Aluminium Compounds, Inorganic-Aluminium Chloride,
2002.
[14] H. Wayne Richardson, Ullman’s Encyclopedia of Industrial Chemistry,
sixth ed., Electronic Release. Copper Compounds—Salts and Basic Salts,
2002.
[24] D.M. Pasquevich, A.M. Caneiro, Thermochim. Acta 156 (2) (1989)
275–283.
[25] F.C. Gennari, D.M. Pasquevich, Thermochim. Acta 284 (1996) 325–339.
[26] A. Roine, Outokumpu HSC chemistry for Windows, in: 93001-ORGT
Version 2. 0, Outokumpu Research Oy Information Service, 1994.
[27] F.J. Alvarez, G. De Micco, A.E. Bohe´, D.M. Pasquevich, IAEA. Technical
Cooperation Regional Project, RLA/4/018, Final Report, March 2005.
[28] Joint Committee for Powder Diffraction Standards, International Center
for Diffraction Data. Powder Diffraction File, 2001.