J Therm Anal Calorim (2010) 100:847–851

DOI 10.1007/s10973-010-0742-8

The role of thermal analysis in detecting impurity levels

during aluminum recycling
Musbah Mahfoud • A. K. Prasada Rao •

Daryoush Emadi

CTAS2009 Special Chapter

Ó Akadémiai Kiadó, Budapest, Hungary 2010

Abstract Recycling of aluminum scrap has gained applications. Aluminum has been recognized as one of the
interest owing to its economic and ecological benefits. best candidate materials for various applications by dif-
Unfortunately, during the collection of scrap from a mixer ferent sectors such as automotive, construction, aerospace,
of junk from various sectors it is difficult to ensure that the etc. The increasing demand for aluminum-based products
recycled alloy has the same chemical composition as that and further globalization of the aluminum industry have
of already existing commercial alloys. Consequently, some contributed significantly to the higher consumption of
of the alloying elements become trace/tramp elements in aluminum scrap for re-production of aluminum alloys.
the recycled alloy. Therefore, in order to obtain high per- Today, a large amount of new aluminum products are made
formance recycled alloys, controlling the impurity levels of by recycled alloys. This represents a growing ‘‘energy
the aluminum melt is of vital importance. Normally, bank’’ of aluminum available for recycling at the end of
computer aided cooling curve analysis (CA-CCA) is used components’ lives, and thus recycling has become a major
to find the relationship between cooling curve parameters, issue. The future growth offers an opportunity for new
melt treatments, alloy composition, and properties. In the recycling technologies and practices to maximize scrap
present study, the first differential thermal analysis (DTA) quality; improve efficiency and reduce cost. One major
approach has been used to detect and quantify impurity hurdle in the direct use of recycled aluminum for new
elements in scrap-like liquid aluminum alloys. applications is the level of impurities present in the recycled
alloy.
Keywords Aluminum recycling
Thermal analysis
Iron is one of the most undesirable impurities, which is
Trace elements
Computer-aided cooling curve analysis
very common in recycled aluminum alloys [1]. It is well
Solidification known that an excess amount of iron in aluminum–silicon
casting alloys can lead to the formation of the Al5FeSi
intermetallic (the b-phase), along with other Al–Fe–Si type
Introduction phases, which can be very detrimental to the mechanical
properties of the final cast part. Khalifa et al. [2] and Rilvin
Among the light metal family, aluminum has been et al. [3] had reported various intermetallics formed in the
acquiring increasing significance for the past few decades Fe containing Al–Si alloys, among them ternary AlFeSi
due to its excellent properties and diversified range of intermetallics are prominent. Further these AlFeSi inter-
metallics are found in two different morphologies for the
same stoichiometry, such as a and b. Figure 1a shows a
M. Mahfoud
A. K. Prasada Rao
Department of Mechanical Engineering, College of the North typical microstructure of a hypoeutectic Al–Si alloy with a
Atlantic-Qatar, Doha, Qatar coarse platelet of Al5FeSi (b-phase). However, the platelet
morphology of Al5FeSi phase can be modified to chinese-
D. Emadi (&)
script like morphology (primary a) with the addition of Mn
Department of Mechanical and Industrial Engineering,
Qatar University, Doha, Qatar to the melt (Fig. 1b) [4]. Narayanan et al. [4] have shown
e-mail: demadi@qu.edu.qa that the cooling rate has a significant role to play in the

123
848 M. Mahfoud et al.

alloy composition, and properties. CA-CCA also readily

incorporates procedures to evaluate parameters such as the
total latent heat, fraction solid, etc., for multi-component
alloys from the cooling curve [10, 12–14]. On the other
hand, heat-pipe technology developed by one of the present
authors has been reported to improve the efficiency of the
conventional CCA by controlling the rate of solidification.
The detailed description of heat-pipe technology has been
Fig. 1 Scanning electron micrographs showing the microstructure of
reported elsewhere [1, 15].
Fe containing A319 alloy a without Mn and b with Mn addition [4]
A319 alloy is used in this study, which is mostly used in
power train applications, such as engine blocks and cylinder
formation of primary a or b phase intermetallics in the case heads. In the present work, an attempt has been made to
of Fe and Mn-containing hypoeutectic Al–Si alloys. It has evaluate (detect and quantify) the impurity elements that are
also been reported that such modification would enhance present in an A319 alloy melt (iron in particular) by using the
the tensile strength of Al–Si alloys [5]. computer aided cooling curve analysis coupled with a heat-
In addition, some elements such as Sn, Sc, Zr, Ti, Sr, pipe and first derivative approach.
and Sb, which are added to improve the properties of Al
alloys [6–9], eventually become trace/tramp impurities
when these alloys are recycled. Unfortunately, these Experimental details
residual elements, which have been ignored by researchers,
could have a detrimental effect on the properties of the Initially 5.0 kg of A319 alloy (commercial Al alloy with
alloy. The effects of these impurities on the mechanical the chemical composition: 6.0 wt%Si, 3.5 wt%Cu,
properties of recycled aluminum alloys is a function of 1.0 wt%Zn, 0.1 wt%Mg, 0.65 wt%Fe, 0.12 wt%Ti) was
their concentrations, and they need to be monitored and melted in a clay graphite crucible using a pit-type resis-
controlled well in advance prior to casting. tance furnace. Later, specific amounts of Fe as mentioned
It is well known that the solidification process of a metal in Table 1 were added to the melt in the form of an Al–Fe
or an alloy is accompanied by the evolution of heat, the master alloy. Thermal analysis was performed by inserting
magnitude of which depends on the various phases formed the leads of the CA-CCA method, coupled with the heat-
during solidification. Recorded temperature–time data can pipe into the melt and the data were recorded using a data
yield quantitative information about the alloy’s solidifica- acquisition system at a rate of 10 readings per second.
tion behavior. The study of cooling curves has shown that it During this process, only the melt in the heat-pipe graphite
can be used to control and optimize the solidification cup (100 g of aluminum sample) was allowed to solidify,
processes. Solidification curves can be analyzed by various where the actual thermal analysis was conducted (while the
methods. The computer aided cooling curve analysis rest of the melt in the crucible was maintained at a con-
(CA-CCA) method is the most recent technique developed stant temperature of 700 °C). This experiment was repe-
for analyzing solidification behavior of metals and alloys ated for different combinations of Fe concentrations and
[10, 11]. This technique was used for many years in fun- cooling rates. The results thus obtained are reported herein.
damental metallurgical studies and for determining binary
phase diagrams. This technique is based on the analysis of
the cooling curve obtained during solidification. A critical
requirement of this method is the zero curve. The zero
curve is, in principle, the derivative of a cooling curve Table 1 Formation temperature, Tfor, of Al5FeSi as a function of
obtained from the metal cup system, assuming that the cooling rate and Fe content
metal does not undergo any phase transformation. In the wt%Fe added Tfor/°C
case of a freezing alloy, phase transformations do not occur
(dT/dt)a-Si/°C-1s-1
at a single temperature, and therefore, the estimation of a
baseline can be complicated and subjective. Traditional Target Analyzed 0.3 0.7 1
cooling curve analysis (CCA) uses a small cup with 1–2
0.65 0.646 574.2 568.9 –
thermocouples, and use the Newtonian or Fourier analysis
0.80 0.789 578.2 573.7 572.2
method to generate the zero curve needed for further
0.95 0.947 583.9 578.6 576.3
analysis of the cooling curve [10].
1.10 1.071 587.7 583.7 579.6
One of the applications of CCA is finding the relation-
1.25 1.218 592.8 587.8 583.9
ship between cooling curve parameters, melt treatments,

123
The role of thermal analysis in detecting impurity levels 849

Results and discussion (a) 620 3

0.8 wt%Fe
The following section demonstrates a method for deter- 610 2

mining the trace elements present in an aluminum melt

Temperature/°C
using thermal analysis during alloy solidification. 600 1

dT/dt (C/s)
In the conventional practice, the presence of such
impurities is usually detected after the alloy has solidified. 590 0

However, in the present work the detection of impurity

580 –1
elements like Fe can be done online prior to casting with
the aid of CA-CCA. This is achieved by coupling heat-pipe
570 –2
technology (which is an excellent tool for controlling the 572.2
solidification rate) with the conventional cooling curve –3
560
analysis. It is well known that phase transformation during 0 10 20 30 40 50 60 70 80
solidification is accompanied by heat evolution, which Time/s
results in a change in the slope of the cooling curve. As (b) 620 3
small changes are usually difficult to detect on the cooling
1.25 wt%Fe
curve, other features of thermal analysis must be used to 610 2
detect these minor reactions. Differential thermal analysis

Temperature/°C
(DTA) uses the derivative curve to accentuate these heat 600 1

dT/dt (C/s)
effects. If iron levels are high enough in hypoeutectic
Al–Si alloys, then the intermetallic phase Al5FeSi often 590 0

forms as a primary phase before the eutectic. At relatively

580 –1
low levels, the iron reaction is not visible on the temper- 583.9
ature plot, but manifests itself as an inflection. Figure 2 –2
570
shows a thermal analysis curve, which confirms the for-
mation of Al5FeSi intermetallic. In Fig. 2, the first deriv- 560 –3
ative curve shows a sharp peak indicating the formation of 0 10 20 30 40 50 60 70 80

the Al5FeSi phase, although such sharp peak is not Time/s

observed in the Temperature–Time curve.
Fig. 3 Cooling curves of A319 alloy containing (a) 0.8 wt% and
In addition to its ability for detecting the presence of (b) 1.25 wt% of Fe
impurities, the present method can also quantify the
impurity elements present in an aluminum alloy melt. This
functionality has been proved by the thermal analysis of that the formation of the Al5FeSi intermetallic evolves
A319 alloy melt containing iron in various concentrations. heat. In other words, the total amount of heat evolved
The results presented in Fig. 3a, b represent the cooling depends mainly on the total volume of the Al5FeSi inter-
curves obtained using A319 alloy with 0.8 and 1.25 wt% of metallic formed during solidification. Hence, it can be
iron, respectively (for a cooling rate of 1 °C/s). It is known understood that the total heat evolved increases with an
increase in the iron concentration in the melt. This is
indicated by the increase in the formation temperature with
700
an increase in the iron content from 0.8 wt% (Fig. 3a) to
1.25 wt% (Fig. 3b). Interestingly, a shift in the peak of the
Temperature/°C/first derivative

680
660
Temp. first derivative curve is observed upon an increase in the
640
620
iron concentration. This can be attributed to the increase in
600 the activity level of iron in melt. In other words, the
dT/dt
580 Al5FeSi intermetallic phase forms faster in aluminum
560
540
alloys with higher iron contents.
520 Similar experiments were conducted for various com-
500 AlFeSi Needle binations of iron concentrations and cooling rates, as
480 Phase
460
shown in Table 1 and Fig. 4. The latter shows that the
440 formation temperature of the Al5FeSi phase is a linear
0 20 40 60 80 100
function of the iron concentration of the melt. It is inter-
Time/s
esting to note that the present results are well in agreement
Fig. 2 Detection of Fe-intermetallic using first derivative approach with an earlier work by Mackay et al. [16]. It is worth

123
850 M. Mahfoud et al.

600 (a) 650 0.2

Mackay:0.1 °C/s Mg Si(602°C)
595 0.1
Formation temperature/°C
Present work: 0.3 °C/s 2
Present work: 0.7 °C/s

Temperature/°C
Present work: 1 °C/s 600 0
590
dT/dt
–0.1
585
550 –0.2
580 Temp.
–0.3
575 500 –0.4

570 T = 31.05%Fe + 553.99 –0.5

T = 31.76%Fe + 548.36
565 450 –0.6
T = 26.26%Fe + 550.97 0 200 400 600 800 1000 1200 1400 1600
560
0.5 0.65 0.8 0.95 1.1 1.25 1.14 Time/s
wt%Fe 1
(b) 620
610 0.5
Fig. 4 Linear relationship between Fe content and formation 600
temperature of Al5FeSi phase 590
Al Cu(585°C) 0

Temperature/°C
2
580 –0.5

dT/dt (C/s)
570
mentioning that the scattering of data points in the present –1
560
work is comparatively less than that observed in Mackay’s 550 –1.5
work (Fig. 4). This can be attributed to the better control of 540
530 –2
the cooling rate in the heat-pipe technology used in the
520
–2.5
present study. The linear equations developed in Fig. 4 can 510 Formation of Al Cu
2
be used to determine the quantity of iron present in the 500 –3
0 10 20 30 40 50 60 70 80 90 100 110 120
melt. Once the formation temperature of the Al5FeSi phase
is measured for a given cooling rate, the amount of iron Time/s
present in the melt can be easily estimated. Figure 5 shows
Fig. 6 Detection of a Mg2Si and b Al2Cu intermetallics using the
the same values (as in Fig. 4) being superimposed on the first derivative approach
Al–Fe phase diagram reported by Phillips [17, 18] with
6 wt%Si. Ideally, the lines representing the slope of each of b shows the thermal analysis plots of an Al–Si alloy con-
the three data should coincide with the slope of the liquidus taining Mg and Cu, respectively. The temperatures of
line in the phase diagram. However, since phase diagrams formation of Mg2Si and Al2Cu were measured by the DTA
are representative of equilibrium conditions, one would technique for a cooling rate of 1 °C/s. Further, these tem-
expect a higher cooling rate to shift the data below the perature values can be used to compute the Mg and Cu
equilibrium line, as can be seen in Fig. 5. concentrations in the Al melt, once the relationship
Similarly, the present method of thermal analysis can be between the concentrations versus formation temperature is
extended to evaluate other trace/impurity elements present established (following the same method as that followed in
in aluminum alloys while in the liquid state. Similar studies the case of iron).
were done on magnesium (Mg) and copper (Cu). Figure 6a,

620
Conclusions
Al + + liquid
610 The present scenario of aluminum industry demands high
Al + liquid Al + + liquid quality recycled aluminum products owing to several
Temperature/°C

600
economic and environmental aspects. Quantifying the
590
levels of impurity elements in the recycled aluminum
580 alloys would allow manufactures to control their process
570 Al +
parameters accordingly. This enables the foundry man to
+ Si
compensate for the detrimental effects of these impurities,
560 and therefore produce high quality castings from recycled
550 aluminum alloys. However, few methods are presently
0 0.2 0.4 0.6 0.8 1 1.2 1.4 available for improving the melt quality, especially for
wt%Fe
producing sound alloys from recycled aluminum. Thus,
Fig. 5 Correlation of the Fe content and formation temperature of there is a need for developing new methods which can aid
Al5FeSi phase with phase diagram (at Si = 6 wt%) in the improvement of melt quality.

123
The role of thermal analysis in detecting impurity levels 851

One of such methods is the thermal analysis technique 5. Sreeja Kumari SS, Pillai RM, Rajan TPD, Pai BC. The effects of
which has been developed and reported in this work, in individual and combined additions of Be, Mn, Ca and Sr on the
solidification, structure and mechanical properties of Al–7Si–
particular that of the heat-pipe method. It has been proved 0.3Mg–0.8Fe alloy. Mater Sci Eng. 2007;460-461A:561–73.
that the presence of an impurity element, such as iron, in 6. Prasada Rao AK, Das K, Murty BS, Chakraborty M. Prediction of
the Al alloy melt can be detected and quantified using the grain size of Al–7Si alloy by neural networks. Mater Lett.
thermal analysis (CCA) method successfully. This method 2008;62(2):273.
7. Prasada Rao AK, Das K, Murty BS, Chakraborty M. Al-Ti-C-Sr
can be further extended to detect and quantify impurities in master alloy - A melt inoculant for simultaneous grain refinement
aluminum alloys prior to casting, especially in the case of and modification of hypo-eutectic Al-Si alloys. J Alloys Compd.
elements like Sn, Sc, Mg, Cu, Ti, Zr, Sb, Zn, etc., which 2009; 480(2):L49–51.
readily form intermetallic phases with aluminum. 8. Mahfoud M, Emadi D. Advances in materials and processing
technologies (AMPT) conference. Manama, Kingdom of Bahrain,
In the present work, it has been found that there is a Nov; 2008.
linear relationship between the Fe content and formation 9. Mahfoud M, Emadi D. 2nd International aluminum casting
temperature of Al5FeSi phase with phase diagram (at conference & exhibition, Moscow, Russia, April 1–3; 2009.
Si = 6 wt%). This line almost has the same slope as the 10. Emadi D, Whiting LV. Determination of solidification charac-
teristics of Al-Si alloys by thermal analysis. AFS Trans. 2002;
liquidus line of the Al–Fe phase diagram (at Si = 6 wt%), 110(1):285–96
which enables the prediction of the formation of Al5FeSi 11. Fornaro O, Palacio HA. Study of dilute Al–Cu solidification by
phase prior to casting well in advance. cooling curve analysis J Mater Sci. 2009; 44(16):4342–4347.
12. Emadi D. In: Proceedings of the 12th annual technical meeting &
Acknowledgements The authors would like to express their exhibition of the Canadian thermal analysis society (CTAS),
appreciation to Qatar National Research Fund for their financial Mississauga, Canada, 14–15 May 2002; p. 58.
support. The technical assistance of the staff at Qatar University 13. Pelayo G, Sokolowski JH, Lashkari R. A case based reasoning
(Department of Mechanical and Industrial Engineering), McGill aluminium thermal analysis platform for the prediction of W319
University (Department of Materials Engineering) and College of Al cast component characteristics. J Achiev Mfg Eng. 2009;
North the Atlantic-Qatar is very much appreciated. 36(1):7–17.
14. Malekan M, Shabestari SG. Effect of grain refinement on the
dendrite coherency point during solidification of the A319
aluminum alloy. Metall Mater Trans. 2009;40A:3196–203.
References 15. Mahfoud M, Emadi D. Application of heat pipe technology
in thermal analysis of metals. J Therm Anal Calorim. 2005;81:
1. Mahfoud M. Controlled thermal analysis using heat pipe tech- 161–7.
nology. PhD Thesis, McGill University, 1997. 16. Mackay RI. Quantification of iron content in Al–Si foundry
2. Khalifa W, Samuel FH, Gruzleski JE. Nucleation of solid alu- alloys via thermal Analysis. Master’s Thesis, McGill University,
minum on inclusion particles injected into Al–Si–Fe alloys. 1996.
Metall Mater Trans. 2003;34A:807–25. 17. Phillips HWL. Institute of Metals Monograph and Report Series,
3. Rilvin VG, Raynor GV. Critical evaluation of constitution of 1959; No. 25: p. 56.
aluminum–iron–silicon system. Int Metall Rev. 1981;3:133–52. 18. Phillips HW. Annotated equilibrium diagrams of some aluminum
4. Narayanan A, Samuel FH, Gruzleski JE. Crystallization behav- alloy systems. London: The Institute of Metals; 1976.
iour of iron-containing intermetallic compounds in 319 alumin-
ium alloy. Metall Mater Trans. 1994;25A:1761–73.

123