*Article* **Stabilization Mechanism of Semi-Solid Film Simulating the Cell Wall during Fabrication of Aluminum Foam**

#### **Takashi Kuwahara 1,\*, Akira Kaya 1, Taro Osaka 1, Satomi Takamatsu <sup>1</sup> and Shinsuke Suzuki 1,2**


Received: 23 January 2020; Accepted: 26 February 2020; Published: 2 March 2020

**Abstract:** Semi-solid route is a fabrication method of aluminum foam where the melt is thickened by primary crystals. In this study, semi-solid aluminum alloy films were made to observe and evaluate the stabilization mechanism of cell walls in Semi-solid route. Each film was held at different solid fractions and holding times. In lower solid fractions, as the holding time increases, the remaining melt in the films lessens and this could be explained by Poiseuille flow. However, the decreasing tendency of the remaining melt in the films lessens as the solid fraction increases. Moreover, when the solid fraction is high, decreasing tendency was not observed. These are because at a certain moment, clogging of primary crystals occurs under the thinnest part of the film and drainage is largely suppressed. Moreover, clogging is occurring in solid fraction of 20–45% under the thinnest part of the film. Moreover, the time to occur clogging becomes earlier as the solid fraction increases.

**Keywords:** porous metal; semi-solid; aluminum foam; primary crystals; drainage; clogging

#### **1. Introduction**

Recent days, improved fuel consumption is required for transportation equipment such as automobiles. Furthermore, passenger safety including the crash safety must be maintained. For this purpose, aluminum foams with many pores dispersed inside are actively investigated. They are ultralight materials and exhibit shock absorbing characteristics called plateau phenomenon when compressed. Additionally, aluminum foam can absorb shocks isotropically [1]. Therefore, it is expected to be applied for transportation equipment [2]. Moreover, they have excellent property in sound insulation equivalent to glass wool [3,4]. Furthermore, they also have heat insulation properties [5]. Furthermore, it can be applied for architecture material [6].

However, there is a problem of drainage during fabricating aluminum foam which coarsens pores and thus, deteriorates the mechanical properties. Usually, aluminum foams are obtained by solidifying the aluminum alloy melt with pores generated inside. During holding of the foam before solidification, liquid flows downwards in a cell wall (a film between pores) due to the gravity. This phenomenon is called drainage. Due to the remarkable collapse of cell walls, pore morphology becomes uneven. Moreover, unfoamed parts would appear at lower part of the foam. However, by suppressing the drainage by thickening the melt, foam with uniformed pore morphology could be obtained [7]. Fabrication of foams with uniform pores becomes possible by stabilizing the cell walls through thickening.

There are several routes to fabricate aluminum foams. Melt-route, shown in Figure 1 is the most known route to fabricate aluminum foam. First, the melt is thickened by oxide particles. Then, the melt is stirred with adding blowing agent (TiH2) included. Finally, an aluminum foam can be obtained by foaming and water cooling. However, our group is fabricating foams in the Semi-solid route shown in Figure 1. The fabrication method is almost the same as the Melt route except for thickening the melt by primary crystals existing in semi-solid state [8]. In the Semi-solid route, the adding of thickener, such as ceramic particles and Ca, can be excluded. Therefore, foam with less impurities can be obtained. Moreover, foams obtained by the Semi-solid route are thought to have higher compressive yield stress than that obtained by the Melt route [9].

**Figure 1.** Fabrication methods of aluminum foam (Melt route and Semi-solid route).

Jin et al. evaluated the Melt route with SiC particles as a thickener and elucidated that the parameters to fabricate stabilized foams (foams with stabilized cell walls) are particle size and particle fraction. Similarly, they expressed the parameter map for fabricating of stabilized foams [10]. Moreover, Heim et al. fabricated foams with various types of thickener including primary crystal and classified them as foamable, partially foamable, or unfoamable. They also expressed that the parameter map proposed by Jin et al. could be adapted to various types of thickener [11].

Moreover, in the Melt route, Heim et al. expressed that some thickener particles are covered by oxide skin of cell wall and interact and builds bridge with other particles resulting to prevent drainage [12]. This is thought to be the stabilization mechanism of cell wall in the Melt route. However, the stabilization mechanism of cell wall in the Semi-solid route has not been evaluated. Furthermore, size of primary crystals is quite large and out of the stabilization region in the parameter map proposed by Jin et al. [10]. Likewise, primary crystals were classified as partially foamable or unfoamable by Heim et al. [11]. Despite, our group found it possible to fabricate stabilized foams in the Semi-solid route out of the stabilization region in the map. Therefore, cell wall of Semi-solid route is thought to have different stabilization mechanism from Melt route. Therefore, elucidation of stabilization mechanism of cell wall of Semi-solid route is required.

However, it is difficult to evaluate the cell walls since the size and shape of cell walls are uneven. As a solution, Heim et al. simulated the ideal shape of the cell wall using aluminum alloy film [13]. In this method, we can evaluate the cell wall in higher comparability. Therefore, the objective of this study is to elucidate the stabilization mechanism of cell wall of the Semi-solid route using aluminum alloy film.

#### **2. Materials and Methods**

#### *2.1. Sample Material*

Sample material used in this study was Al-6.4mass%Si. Since the phase diagram of Al-Si is simple and primary crystals of hypoeutectic alloy are easy to observe, this material is suitable for evaluation of stabilization mechanism. The Al-6.4mass%Si ingots were prepared by mixing pure aluminum with Al-25mass%Si. The compositions of the pure aluminum and the Al-25mass%Si ingots used in this study are shown in Tables 1 and 2, respectively.


**Table 1.** Composition of pure aluminum.

#### *2.2. Forming of Aluminum Alloy Film*

Aluminum alloy films which are simulating the cell walls of aluminum alloy foam were formed to evaluate the stabilization mechanism of cell wall in higher comparability. In the Semi-solid route, the melt was thickened in semi-solid state as shown in Figure 1. Therefore, aluminum alloy films were also formed in the semi-solid state to observe the cell walls in the Semi-solid route. The solid fractions of the melt of Al-6.4mass%Si were obtained accurately using Thermo-Calc 2019a (Itochu Techno-Solutions Corporation, Tokyo, Japan) as shown in Figure 2.

**Figure 2.** Al-Si binary phase diagram. The values of solid fraction were calculated by Thermo-Calc.

All of processes to form aluminum alloy films were done inside a chamber under pressure condition of 5000 Pa of air to suppress heat dissipation from the film during the retaining process described later. To obtain this condition, pressure was adjusted by vacuuming continuously while controlling the leaking rate with valve. To form the aluminum alloy film, Al-6.4mass%Si was completely molten in the crucible. Then, the melt was slowly cooled to the semi-solid temperature. The temperature was measured by K-type thermocouple coated with heat resistant inorganic adhesive (Toagosei Co. Ltd., Tokyo, Japan) soaked in the melt. Then, the melt was stirred for 60 s at 900 rpm so as to round the shape of primary crystals [14]. Before this process, thermocouple was taken out of melt to make the stirring possible. Since the stirring generates the temperature of melt to change, the thermocouple was soaked again in the melt and the temperature was adjusted again to semi-solid temperature.

After the semi-solid temperature reached the setting temperature, a wire shown in Figure 3a was dipped into the melt. This wire was shaped so as to form a liquid film between the rings when pulled up from the melt. The design of this wire originates from a previous research [13]. In this study, it was adopted to study the semi-solid densification phenomena. The material of the wire was stainless steel SUS304 and the drainage channel was provided for the melt to flow down from the film and reproduce the drainage. The ratio of distance between the rings versus its diameter was one versus three following the previous study to form a film with a continuous curvature [13]. After pulling up the film from the melt, the film was retained over the melt with drainage channels contacted with the melt to reproduce the drainage [13]. This process simulates the progression of drainage which causes collapse of cell walls in aluminum form. The holding time was changed in each film. Finally, the aluminum alloy film was obtained by pulling up the film at 5 mm/s and rapid cooling by dipping into a melt pool of low melting point alloy U78 (Osaka Asahi Co. Ltd., Osaka, Japan) held at 150–200 ◦C in the chamber. The schematic of the experimental procedure is shown in Figure 4.

**Figure 3.** Schematic of wire and film: (**a**) size of wire; (**b**) cutting of film; (**c**) measurement of initial cross-sectional area; and (**d**) measurement of cross-sectional area.

**Figure 4.** Schematic of forming liquid film of aluminum alloy.

Holding times of the aluminum alloy films were 0, 30, 60, 90, and 120 s. However, since it took a little time to move film to cool down, actual time of drainage was longer than the time designated. The solid fractions of melt were 5%, 11%, and 18.5%. Aluminum alloy film was formed for each combination of holding time and solid fraction. For solid fraction of 5%, holding time of 15 s and 45 s were also obtained for further consideration.

#### *2.3. Observation of Aluminum Alloy Film*

The formed film was first rinsed inside with the boiled water to remove U78 used for rapid cooling. Then, formed films were cut in the middle as shown in Figure 3b and the cross sections were observed by optical microscope. Primary crystals with a diameter larger than 200 μm, which seemed to have existed in a semi-solid state, were colored on micrographs with using an image processing software GIMP 2.10.4 (The GIMP Development Team, Berkeley, CA, USA). Then, the cross-sectional area of the films and the area of colored primary crystals were measured with using an image processing software WinROOFTM 6.4.0 (Mitani corporation, Fukui, Japan).

The cross-sectional area was measured to evaluate the change of melt quantity during the holding time. Furthermore, comparison of cross-sectional area was done by normalized cross-sectional area, as a parameter of the area of the remained melt compared with the initial one. To obtain the normalized cross-sectional area, initial cross-sectional area A0 is defined as area surrounded with wires (Figure 3c). Then the normalized cross-sectional area was obtained through dividing the cross-sectional area A (Figure 3d) by the initial cross-sectional area. However, small deformation occurred on wires because of the flow stress of the semi-solid state. Therefore, the wires of the actual results were not aligned as schematic in Figure 3a. Likewise, this tendency is remarkable in the lengths between the rings while the lengths from the diameter of rings show relatively close value. Therefore, the initial cross-sectional

area was obtained by calculating the area of trapezoid. The upper and lower length between the rings were used as upper and lower bases of the trapezoid. The average diameter of rings was used as its height.

#### **3. Results**

#### *3.1. Temporal Change of Cross Sectional Area*

The plots in Figure 5 show the temporal change of the normalized cross-sectional area obtained from the experimental results. From Figure 5, the normalized cross-sectional area decreases as the holding time increases at lower solid fractions. Additionally, decreasing tendency of the normalized cross-sectional area lessens as the solid fraction increases. Moreover, at high solid fraction this decreasing tendency was not observed.

**Figure 5.** Temporal change of normalized cross-sectional area: (**a**) solid fraction *f* s = 5%; (**b**) *f* s = 11%; and (**c**) *f* s = 18.5%.

#### *3.2. Observation of Primary Crystals*

Figure 6b shows the cross-sectional photomicrograph of an aluminum alloy film shown in Figure 6a. In Figure 6c, primary crystals that seem to have been existing in the semi-solid state are colored as expressed in Section 2.3. Most of them are observed in the lower part of the film in the micrograph and less primary crystals were observed in the drainage channel than in the ring. Therefore, primary crystals seem to have clogged in the lower part of the film. However, only for the film for *f* <sup>s</sup> = 11%, *t* = 0 s, primary crystals that seem to have been existing from semi-solid state were not observed (Figure 7). This is an exception. In this experiment, the part with a small solid fraction might have been pulled up for some reason, such as inhomogeneous solid fraction in the crucible.

**Figure 6.** Cross section of aluminum alloy film (solid fraction *f* s = 18.5%, holding time *t* = 120 s): (**a**) cross section in film; (**b**) photomicrograph; and (**c**) primary crystals colored on photomicrograph.

**Figure 7.** Photomicrograph of cross section of aluminum alloy film (solid fraction *f* s = 11%, holding time *t* = 0 s).

#### **4. Discussion**

#### *4.1. Suppression of Drainage by Thickening*

As descripted in Introduction, thickening of melt stabilizes the cell walls and could homogenize the pore morphology. Increase of viscosity by thickening suppresses the liquid flow in aluminum alloy film. This is thought to be the stabilization mechanism of cell wall. For confirmation, we derived a model formula of melt flow and compared it with the experimental results. The model formula describes relationship between the holding time and the normalized cross-sectional area of the film.

#### 4.1.1. Derivation of Model Formula

Figure 8 shows the change in geometry of the aluminum alloy film. Heim et. al. approximated the cross-sectional shape of the pulled-up film as a rectangle [12]. Then, the film will start to curve as the drainage progresses. The curvature radius of the film is decided by approximation to an arc defined by three points; the top and bottom of the wire and the middle of the film. The middle point of the film should shrink as the drainage progresses. Therefore, the curvature radius of the film should change continuously. The depth of the film *L* would not change in this model. This is because from observation of films including Figure 6, change in depth of film is small compared to width of film *x0*. This may be because it is difficult to make curvature radius smaller when the original curvature radius is small. In this situation, the relationship between the cross-sectional area *A* of film and width of film *x*t is shown in Equation (1). (The discussion is shown in Appendix A).

$$A = h\mathbf{x}0 - \frac{1}{4\Delta\mathbf{x}} \left\{ \frac{\pi}{2\Delta\mathbf{x}} \left( h^2 + \Delta\mathbf{x}^2 \right)^2 \frac{\partial}{360^\circ} - \left( h^2 + \Delta\mathbf{x}^2 \right) h \right\} \tag{1}$$

**Figure 8.** Geometry model of aluminum alloy film.

Here, θ is curvature radius of the film, *h* is height of rectangle, *x*<sup>0</sup> is width of rectangle, and *x*<sup>t</sup> is width of middle point of rectangle which would change continuously. Furthermore, Δ*x* is defined as Equation (2).

$$
\Delta \mathbf{x} = \mathbf{x}\_0 - \mathbf{x}\_t \tag{2}
$$

Moreover, the curvature radius of the film θ can be shown in Equation (3).

$$\theta = \text{Arcsin} \frac{4h(\mathbf{x}\_0 - \mathbf{x}\_t) \left\{ h^2 - (\mathbf{x}\_0 - \mathbf{x}\_t)^2 \right\}}{\left\{ h^2 + (\mathbf{x}\_0 - \mathbf{x}\_t)^2 \right\}^2} \tag{3}$$

However, since the changing parameter is width of film at middle *x*t, Equation (1) could not explain the temporal change of the cross-sectional area of film *A*. Therefore, a fluid dynamic model of film is considered. Brady et. al. considered that liquid (semi-solid) flow in a film by drainage can be described as the Poiseuille flow under condition that the ends of both sides are fixed [15]. The situations are, one way flow, steady flow, and gravity as the only external force, as shown in Figure 9a. Following the Poiseuille flow, width of film at middle *x*t can be shown as Equation (4).

**Figure 9.** Schematics of Poiseuille flow and deformation of the film: (**a**) at *t* = 0 s and (**b**) at *t*.

Here, μ is the viscosity of melt, ρ is density of melt and *g* is the gravitational acceleration. By Equation (4), relationship between the width of film at middle *x*<sup>t</sup> and holding time *t* could be obtained. By substituting the width of film at middle *x*<sup>t</sup> obtained in Equation (4) into Equation (2), relationship between the cross-sectional area of film *A* and holding time *t* could be obtained by Equations (1) and (3). Therefore, a model formula to compare with the result is obtained. Figure 9b shows a schematic model of the phenomenon expressed by this model formula. Correspondingly, in this model, since the mainly considered area is middle of the film, which is assumed that drainage occurs mainly, three-dimensional deformation and drainage from ends of film is not considered.

#### 4.1.2. Comparing Cross Sectional Areas Obtained from the Model with Experimental Results

To compare the model and the experimental results, the model formula was transformed into a dimensionless formula shown in Equation (5). *A0* in Equation (5) is the initial cross-sectional area and can be shown in Equation (6).

$$\frac{A}{A\_0} = 1 - \frac{1}{4h\mathbf{x}\_0 \Delta \mathbf{x}} \left\{ \frac{\pi}{2\Delta \mathbf{x}} \left( h^2 + \Delta \mathbf{x}^2 \right)^2 \frac{\Theta}{360^\circ} - \left( h^2 + \Delta \mathbf{x}^2 \right) h \right\} \tag{5}$$

$$A\_0 = h\mathbf{x}\_0\tag{6}$$

Figure 5 shows the cross-sectional areas obtained from the model with experimental results. According to the experimental data of Al-6.5mass%Si obtained by Moon et. al. using a Searle-type viscometer [16], the viscosity values of 22, 45, and 100 mPa·s were used for 5%, 11%, and 18.5% of solid

fraction, respectively. Furthermore, the density value used was 2.4×103 kg/m3 for every solid fraction. This value was assumed from study done by Kudoh et. al. which measured the density of liquid phase in semi-solid state of Al-2.4mass%Si [17]. However, change of density between 2.3–2.7×10<sup>3</sup> kg/m3 only produces about 3% change in the calculated value of the model formula and has no effect on discussion. At lower solid fractions, experimental results seem to fit model formula. However, fitting becomes worse as the solid fraction increases. Furthermore, for high solid fraction, fitting of model formula to experimental results was not good. Therefore, it is difficult to conclude that only the increase in fluid viscosity is the mechanism to stabilize the film. Therefore, consideration of other factors to suppress the drainage is required.

#### *4.2. Suppression of Drainage by Clogging of Primary Crystals*

As expressed in Section 3.2, primary crystals seem to have been clogging in the lower part of the film. To confirm and to find the clogging part more finely, the distribution of primary crystals in the films was measured in the height direction from the lower wires (Figure 10). In Figure 10, the primary crystals that seem to have been existing from semi-solid state are colored as expressed in Section 2.3. As the primary crystals were found more in the lower part than in the upper part and drainage channel, clogging should have occurred in the lower part under the thinnest part of the film. Thus, this factor effects to suppress drainage more strongly than thickening. Likewise, at solid fraction of 5%, this tendency was seen remarkably in films with long holding time.

**Figure 10.** Primary crystal ratio: (**a**) solid fraction *f* s = 5%, holding time *t* = 120 s; (**b**) *f* s = 11%, *t* = 90 s; and (**c**) *f* s = 18.5%, *t* = 120 s.

To derive this clogging mechanism more quantitatively, the critical solid fraction *f* s\_cr for clogging was evaluated. Areas of primary crystals *Ap* and film *Af* were measured in the region between the thinnest part of the film and the lower wires (Figure 11). In Figure 11, primary crystals that seem to have been existing in semi-solid state are colored in the same way as expressed in Section 2.3. The area of aluminum alloy film is also colored in the same way as expressed in Section 2.3. By dividing the area of primary crystals *Ap* by area of aluminum alloy film *Af*, the critical solid fraction for clogging was obtained. Figure 12 shows the measured results of the critical solid fraction for clogging. In Figure 12, the critical solid fraction for clogging seems not to change in increase of holding time. Similarly, critical solid fraction for clogging is in the range from 20% to 45%. Therefore, clogging seems to be occurring in this region of solid fraction.

**Figure 11.** Evaluations of area of aluminum alloy film and primary crystals.

**Figure 12.** Relationship between critical solid fraction for clogging *f* s\_cr and holding time for initial solid fraction *fs* of 5%, 11%, and 18.5%.

Clogging of thickener has also been reported for the Melt route [13]. However, bridge formation of thickener attributed to oxide film is thought to be the main stabilization mechanism in Melt route [12]. For Semi-solid route, since the size of thickener is large, clogging is possible to occur below the thinnest part of film and it is assumed that film is not required to be thin for the clogging to occur.

#### *4.3. Fluid Flow and Clogging of Primary Crystals*

According to the results in Figure 5 clogging occurs at a certain timing, which becomes earlier with increasing solid fraction. For solid fraction of 18.5%, clogging seems to have occurred right after the pull-up. However, for the lower solid fractions, drainage following the model formula occurs in a pulled-up film until the clogging. Figure 13 which is a qualitative representation based on Figure 5 summarizes the schematic and graph of this phenomenon. The solid lines show the actual change of normalized cross-sectional area *A*/*A*0. The broken lines and curved solid lines show the values obtained from model formula.

**Figure 13.** Clogging time in model formula and schematic of clogging mechanism. The solid lines show the actual change of normalized cross-sectional area *A*/*A*0. The broken lines and curved solid lines show the values obtained from model formula.

#### **5. Conclusions**

The stabilization mechanism of cell wall in Semi-solid route was evaluated using aluminum alloy films. The results can be summarized as follows.


**Author Contributions:** Conceptualization, T.K. and S.S.; methodology, T.K., A.K., T.O. and S.S.; validation, T.K., A.K., T.O. and S.S.; formal analysis, T.K.; investigation, T.K., A.K., T.O. and S.S.; data curation, T.K. and S.S; writing—original draft preparation, T.K.; writing—review and editing, T.K., S.T. and S.S.; visualization, T.K.; supervision, S.S.; project administration, S.S.; funding acquisition, T.K. and S.S. All authors have read and agreed to the published version of the manuscript.

**Funding:** This study was supported by the Grant-in-Aid the Light Metal Educational Foundation.

**Acknowledgments:** The authors thank The Light Metal Educational Foundation for suppling the pure Aluminum ingots used in this study.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **Appendix A**

Here, derivation of geometry model is expressed. As expressed in Section 4.1.1 and Figure 8, a pulled-up film is approximated as a rectangle. Therefore, the cross-sectional area of the film at this point can be expressed as Equation (A1).

$$A\_0 = h\mathbf{x}\_0\tag{A1}$$

Here, *A*<sup>0</sup> is the initial cross-sectional area, *h* is height of rectangle, and *x*<sup>0</sup> is width of rectangle. Then, the side wall of the film will start to become a curved surface as the drainage progresses. At this point cross section of the film can be shown as Figure A1. Here, θ is the curvature radius of the film, *x*<sup>t</sup> is the width of middle point of rectangle which would change continuously.

**Figure A1.** Coordinate axes of cross section of aluminum alloy film during progress of drainage.

Figure A1 also shows the coordinate axes of aluminum alloy film. Rectangle inside chain line shows the initial cross section. To obtain the cross-sectional area *A* as a formula, area of shaded part *S* in Figure A1 is obtained as follows. The equation of a circle with this curvature can be shown as Equation (A2) since the circle is known to pass through three dark points. Therefore, the radius of the circle can be shown as Equation (A3). Middle point of circle is expressed as light point in Figure A1.

$$\left\{\mathbf{x} - \frac{\left(\mathbf{x}\_0 - \mathbf{x}\_t\right)^2 - h^2}{4\left(\mathbf{x}\_0 - \mathbf{x}\_t\right)}\right\}^2 + \left(y - \frac{h}{2}\right)^2 = \frac{1}{16\left(\mathbf{x}\_0 - \mathbf{x}\_t\right)^2} \left\{\left(\mathbf{x}\_0 - \mathbf{x}\_t\right)^2 + h^2\right\}^2\tag{A2}$$

$$r = \frac{1}{4(x\_0 - x\_t)} \left\{ (x\_0 - x\_t)^2 + h^2 \right\} \tag{A3}$$

Then the area of the light gray triangle *S*' in Figure A1 can be obtained as Equation (A4) from Heron's formula using three sides of triangle.

$$S' = \frac{h}{8(\mathbf{x}\_0 - \mathbf{x}\_t)} \left\{ h^2 - (\mathbf{x}\_0 - \mathbf{x}\_t)^2 \right\} \tag{A4}$$

Since the area of triangle can also be shown as Equation (A5), the curvature radius θ can be obtained as Equation (A6) from Equations (A4) and (A5).

$$S' = \frac{1}{2}r^2 \sin \theta \tag{A5}$$

$$\theta = \text{Arcsin} \frac{4h(\mathbf{x}\_0 - \mathbf{x}\_t) \left| h^2 - \left( \mathbf{x}\_0 - \mathbf{x}\_t \right)^2 \right|}{\left\{ h^2 + \left( \mathbf{x}\_0 - \mathbf{x}\_t \right)^2 \right\}^2} \tag{A6}$$

From the curvature radius, the area of the sector composed of the shaded part and the light gray triangle can be obtained as Equation (A7). Then, by subtracting Equation (A4) from Equation (A7), the area of the shaded part can be obtained as Equation (A8). In the equations, *S* is the shaded part which shows the area of outer part the of curved film.

$$\left(\mathbf{S} + \mathbf{S}'\right) = \frac{\pi}{16\left(\mathbf{x}\_0 - \mathbf{x}\_t\right)^2} \left| h^2 + \left(\mathbf{x}\_0 - \mathbf{x}\_t\right)^2 \right|^2 \frac{\Theta}{360^\circ} \tag{A7}$$

$$S = \frac{1}{8(\mathbf{x}\_0 - \mathbf{x}\_l)} \left[ \frac{\pi}{2(\mathbf{x}\_0 - \mathbf{x}\_l)} \left\{ h^2 + (\mathbf{x}\_0 - \mathbf{x}\_l)^2 \right\}^2 \frac{\partial}{\partial 60^\circ} - \left\{ h^2 - (\mathbf{x}\_0 - \mathbf{x}\_l)^2 \right\} \mathbf{h} \right] \tag{A8}$$

Since there are curved parts at both side of film, the total area of the outer part of the curved film can be shown as Equation (A9).

$$2S = \frac{1}{4(\mathbf{x}\_0 - \mathbf{x}\_l)} \left| \frac{\pi}{2(\mathbf{x}\_0 - \mathbf{x}\_l)} \left\{ h^2 + (\mathbf{x}\_0 - \mathbf{x}\_l)^2 \right\}^2 \frac{\partial}{360^\circ} - \left\{ h^2 - (\mathbf{x}\_0 - \mathbf{x}\_l)^2 \right\} \mathbf{h} \right| \tag{A9}$$

Finally, by subtracting Equation (A9) from Equation (A1), the cross-sectional area of the film with a curvature can be shown as Equation (A10) with Δ*x* defined as Equation (A11).

$$A = A\_0 - 2\mathbf{S} = h\mathbf{x}\_0 - \frac{1}{4\Delta\mathbf{x}} \left\{ \frac{\pi}{2\Delta\mathbf{x}} \left( \mathbf{h}^2 + \Delta\mathbf{x}^2 \right)^2 \frac{\Theta}{360^\circ} - \left( \mathbf{h}^2 + \Delta\mathbf{x}^2 \right) \mathbf{h} \right\} \tag{A10}$$

$$
\Delta \mathbf{x} = \mathbf{x}\_0 - \mathbf{x}\_\mathbf{t} \tag{A11}
$$

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **A-242 Aluminium Alloy Foams Manufacture from the Recycling of Beverage Cans**

#### **Nallely Montserrat Trejo Rivera \*, Jesús Torres Torres and Alfredo Flores Valdés**

Centro de Investigación y de Estudios Avanzados del Instituto Politécnico Nacional, Unidad Saltillo, Avenida Industria Metalúrgica 1062, Parque Industrial Saltillo-Ramos Arizpe, 25900 Ramos Arizpe, Coahuila, Mexico; jesus.torres@cinvestav.edu.mx (J.T.T.); alfredo.flores@cinvestav.edu.mx (A.F.V.) **\*** Correspondence: nallely.trejo@cinvestav.edu.mx; Tel.: +52-844-438-9600

Received: 5 December 2018; Accepted: 11 January 2019; Published: 16 January 2019

**Abstract:** This paper presents and discusses a methodology implemented to study the process of the preparation of aluminium alloy foams using the alloy A-242, beginning from the recycling of secondary aluminium obtained from beverage cans. The foams are prepared by a melting process by adding 0.50 wt.% calcium to the A-242 aluminium alloy with the aim to change its viscosity in the molten state. To obtain the foam, titanium hydride is added in different concentrations (0.50 wt.%, 0.75 wt.%, and 1.00 wt.%) and at different temperatures (923, 948 K, and 973 K) while the foaming time is kept constant at 30 s. For a set of experimental parameter values, aluminium alloy foams with the average relative density of 0.12 were obtained and had an 88.22% average porosity. In this way, it is possible to state that the preparation of aluminium alloy foams A-242 processed from the recycling of cans is possible, with characteristics and properties similar to those obtained using commercial-purity metals.

**Keywords:** aluminium alloy foam; recycling; beverage cans; direct foaming method; A-242 alloy

#### **1. Introduction**

Aluminium and its alloys are highly recyclable metals that are easily used for the preparation of several specific alloys, and there are claims that the practice of recycling aluminium allows to reduce pollution and contributes to saving electrical energy as compared to the primary aluminium obtaining process. The recycling process is suitable as aluminium cans from discarded beverages are composed mainly of aluminium alloys, so any separation methods for other materials are not required [1].

Recycling is desirable when the environmental and economic implications of their reintegration do not exceed the limits of their primary production [2]. Aluminium is used in present times like no other metal, together with steel, so its production is continuously growing, with an average growth of 3.7% per annum [3]. Its physical properties make it an ideal candidate for a range of applications in industries such as packaging, transportation, construction, and aerospace, among others [4]. However, secondary aluminium production faces problems with quality losses (when the purity of the aluminium produced is lower than the input material, e.g., by adding alloying elements during re-melting) and dilution losses (addition of primary aluminium during melting to dilute the concentration of residual elements that cannot be refined), accumulation of impurities, and unlimited options for purification of the molten mass [5–8]. Nevertheless, several studies have addressed the fact that aluminium recycling needs no more than 5% of the energy needed for its primary production, so it presents a real opportunity to reduce environmental impacts if managed in a sustainable way [9,10].

On the other hand, it is well known that for the 2XX series of alloys, copper is the main alloying element, so these types of alloys are thermally treatable. Some of the properties that they present are good ductility, high tensile strength (275 MPa) and compressive yield strength (235 MPa). During solidification, the main alloying elements form intermetallic phases which precipitate in different morphologies, giving rise to various properties. Copper is slightly soluble in aluminium at room temperature (0.2%); however, increasing the temperature (821 K) can dissolve up to 5.65% copper. During the solidification process of the A-242 aluminium alloy, the intermetallic compound Al2Cu is formed as a result of a eutectic reaction, improving resistance and hardness in the as-cast condition. With high copper content (4 to 6%), the aluminium alloy responds strongly to heat treatments. On the other hand, nickel forms the intermetallic Al3Ni with aluminium on solidification, which improves the properties at high temperatures, also reducing the coefficient of thermal expansion. This alloy is used in the manufacture of cylinder heads for motorcycles and pistons [11]. Table 1 shows the chemical composition of the A-242 alloy [12].


**Table 1.** Chemical composition of A-242 aluminium alloy (wt.%).

The automotive industry is constantly searching for innovation in its products, giving the opportunity to develop and innovate new materials which allow a better fuel efficiency and passenger safety [13]. Aluminium foams are an attractive alternative for this purpose because their properties are ideal for this purpose as its lightness and ability to absorb impact energy.

Metal foams are light materials in which a gas dispersed in a metal matrix occupies between 50 and 90% of the total volume, obtaining low density values (0.3–0.8 g/cm3). They are characterized by a combination of mechanical and physical properties resulting from both the porous structure and the characteristics of the metal from which they were manufactured [14,15]. Depending on the production method, the foam structure is more or less homogeneous and comprises different characteristic features that determine its properties and, therefore, the fields of application [16,17].

However, less common and familiar are applications and products based on metal foams. The reason is that they are still not widespread, although they have a very high potential, and a large number of applications already exist on the market. Some reviews about the applications of metallic foams are available in the literature [16–22]; however, in recent years, new application fields have emerged, and not all are considered commercially relevant.

The main applications of metal foams can be grouped into structural and functional applications, and are based on several excellent properties of the material [17]. Structural applications take advantage of the light-weight and specific mechanical properties of metal foams; functional applications are based on a special functionality, i.e., a large open area in combination with very good thermal or electrical conductivity for heat dissipation or as electrode for batteries, respectively [16].

There are a large number of manufacturing methods for metal foams. The closed porosity metal foams are commonly fabricated by direct and indirect foaming methods, such as the melting (ML) and powder metallurgical (PM) methods [16,17], respectively. These methods are already described in the literature [16,17,23–26].

Shinko Wire provided the first commercial production of metal foams in the late 1980s. In this process, calcium is first added to an aluminium melt and stirred in air to produce oxides and raise its viscosity. Subsequently, the powder of the blowing agent (TiH2) is dispersed quickly into the melt by stirring, decomposing into gaseous hydrogen and titanium at the melt temperature. The melt starts then foaming inside the crucible or inside a mold. Finally, it is cooled down to a big block and usually sliced into plates of the desired thickness. This type of metallic foam and the corresponding process is called Alporas, which was patented in America in 1987 [16,27].

Another process to manufacture aluminium alloy foams melt was invented in the early 1990s by Alcan International Limited in Montreal (QC, Canada) [28], and Norsk Hydro (Oslo, Norway). In this process, the melt needs to be prepared with ceramic particles, such as silicon carbide or aluminium oxides, ranging from 5 to 20 vol.% to increase the viscosity of the melt and to stabilize the liquid cell walls. Subsequently, air bubbles are injected and, at the same time, dispersed into the melt using

rotating impellers. The bubbles rise to the top of the melt, where they are collected in a liquid foam and start solidifying after leaving the furnace, where the foam can be continuously drawn off using a conveyor belt. This method is used nowadays by Cymat (Mississauga, ON, Canada) to produce foam panels or to fill molds with foams that do not need further processing [16].

In the 1950s, Allen et al. patented a method for foaming metal [29]. This route consists of an indirect foaming of solid precursors by heating. The precursor is produced by mixing aluminium powders with the corresponding alloying elements and a blowing agent, typically 0.5 to 1.0 wt.% of TiH2. Once the powder mixture is prepared, it is consolidated and sintered by extrusion, uniaxial compaction or rolling to yield a foamable precursor. By the heating of the precursors, the matrix starts melting, and the gas of the blowing agent nucleates. In the course of the temperature increasing, hydrogen production increases, and the gas diffuses to the nucleated pores, letting them grow into big bubbles and expanding the foam. The resident oxides in the metal powders (usually 0.5 to 1%) provide the stability of the foam during the holding time in the liquid state [30]. After several minutes, the foam development is fulfilled, and the foamed metal structure can be conserved by temperature reduction, leading to foam solidification [16].

Applications of metal foams are strongly linked to the properties that such kinds of materials can offer and especially to those that are excellent or even unique. Some of the properties are obviously mainly related to those of the matrix metal itself, e.g., elasticity, temperature or corrosion resistance, etc., while others appear only in combination with the cellular structure, e.g., low density, large surface area or damping [16].

The mechanical properties of metal foams are of course correlated to the ones of the corresponding bulk metal, but in a specific manner. The dominating factors here are the density and the structure itself. The foam structure is obviously the characteristic feature of a foam. Mechanical properties depend mainly on the density but are also influenced by the quality of the cellular structure in the sense of cell connectivity, cell roundness and diameter distribution, fraction of the solid contained in the cell nodes, edges or the cell faces [16,26,31–36].

The main objective of this work was the preparation A-242 aluminium alloy foams, obtained by adjusting the chemical composition of secondary aluminium from the recycling of beverage cans. On the other hand, taking into account that during solidification of the aluminium alloy foam, the precipitation of various intermetallic compounds occurs (i.e., Al2Cu, Al3Ni, Al9FeNi, Al2CuMg, Ti, Al3Ti), it was interesting to study the effects of the processing temperature and the content of titanium hydride used on the formation in the aluminium alloy foams and mechanical behavior during the compression test.

#### **2. Materials and Methods**

#### *2.1. Preparation of Aluminium Alloy Foam*

The cans are composed mainly of three different alloys, i.e., the body corresponds to the A-3004 alloy, the lid to the A-5182 alloy, and the seal to the A-5082 alloy [37]. Once melted, the composition of the obtained alloy is similar to that described in Table 2. Compacted cans were melted in a gas-fired furnace containing a silicon carbide crucible with a capacity of 60 kg. Once the temperature of 1023 K was attained, a flux was added to the molten bath to remove impurities from the alloy.

Once the beverage cans were melted, the chemical composition was adjusted with additions of pure copper (4 wt.%) and electrolytic nickel (2 wt.%) at a temperature of 1293 K to bring the mixture to the A-242 alloy specification. The amounts of alloying elements to be added were calculated using the following equation:

$$X = \frac{(P\mathbf{c})(\mathbb{C})}{100\%},$$

where *X* is the amount, in grams, of the alloying element to be added (Cu, Ni); *P*c is the weight of the load as melted (g); and *C* is the difference between the weight percentage necessary to attain the composition and the initial one. Table 3 shows the final chemical composition of the A-242 aluminium alloy obtained from the adjustment of the chemical composition after the melting step.

**Table 2.** Chemical compositions of the alloys contained in the aluminium cans and the final composition of the fused alloy (base alloy) in wt.%.


**Table 3.** Chemical composition of the A-242 aluminium alloy obtained after the melting and composition adjustment steps.


The chemical compositions of the secondary aluminium and of the A-242 aluminium alloy were determined using a SpectroLAB spark emission spectrometer (Spectro Inc., Kleve, Germany).

To determine the melting temperature range of the A-242 alloy during heating, a Perkin Elmer Differential Thermal Analyzer 7 (Seiko Instruments Inc, Chiba, Japan) was used. Three tests were carried out at a heating speed of 10 ◦C/min under an argon atmosphere. Figure 1 shows a differential thermal analysis (DTA) pattern for the obtained A-242 aluminium alloy. This shows an endothermic event in the range from 869 to 931 K, related to the starting temperatures of the melting of the A-242 aluminium alloy.

**Figure 1.** DTA pattern for the A-242 aluminium alloy.

The next step was the preparation of A-242 aluminium alloy foams. At this step, the effects of the content of the foaming agent and the foaming temperature on the mechanical properties and microstructure of the aluminium alloy foams were investigated. Table 4 presents the parameters and their values studied.


**Table 4.** Values selected for the preparation of A-242 aluminium alloy foams (foaming time is constant for 30 s).

The A-242 aluminium alloy foams were prepared in an electrical resistance furnace equipped with a mechanical agitator using a bipartite steel mold. Once the constant temperature of 923 K was attained, 0.50 wt.% Ca was added, and the mixture was shaken for 5 min at a speed of 1500 rpm in order to modify the viscosity of the aluminium alloy. It is worth mentioning that the inclined plane technique was used to determine the viscosity value attained after this addition. We found that, under the conditions imposed, the aluminium alloy developed a viscosity of 0.196 Pa·s—enough to carry out the foaming process. The foaming agent was added according to the conditions depicted in Table 4, where the TiH2 thermally decomposes, releasing hydrogen [38]. The release of this gas gives rise to the formation of bubbles inside the metal held in a semisolid state. The TiH2 was allowed to react for 30 s with constant agitation at 3000 rpm. After the foaming time had elapsed, the mold was removed from the furnace, allowing the prepared aluminium alloy foam to solidify to room temperature (298 K). When the aluminium alloy solidifies, the bubbles are trapped in the metal, giving rise to the formation of pores inside the alloy. The aluminium alloy foams obtained are cylindrical in shape, an average of 118.06 mm in height, and with a diameter of 75 mm.

#### *2.2. Morphological Characterization*

The expansion of the aluminium alloy only takes place in the direction of the height. Therefore, Equation (2) [39] is used to find the linear expansion of the aluminium alloy as a function of the heights of the aluminium alloy foam and of the molten metal in the mold:

$$
\alpha\_{LE} = \frac{h\_1 - h\_2}{h\_2} \times 100\% \tag{2}
$$

where *αLE* is the linear expansion of the foam, *h*<sup>1</sup> is the height of the foam, and *h*<sup>2</sup> is the height of the molten metal in the mold.

The density of the aluminium alloy foams was determined using Equation (3), where the samples used were in the form of cubes made by cutting the aluminium alloy foam. Each sample was weighed using a digital apparatus obtaining the mass of the specimen (*m*) expressed in grams. The dimensions of the samples were also measured in order to calculate their volume (*V*).

$$
\rho = \frac{m}{V} \tag{3}
$$

The relative density (*ρ\**) of the samples was estimated using Equation (4), where *ρ* corresponds to the density of the aluminium alloy foams obtained from Equation (3), and *ρ<sup>s</sup>* for which the density of A-242 aluminium alloys (2.823 g/cm3) [11] was used.

$$
\rho^\* = \frac{\rho}{\rho\_s} \tag{4}
$$

The percentage of porosity was determined using Equation (5) [39]:

$$P = \left(1 - \frac{\rho}{\rho\_s}\right) \times 100\tag{5}$$

where *P* is the percentage of porosity of the aluminium alloy foams.

To find the pore diameter, the photomicrographs obtained from the stereographic microscope were used. Image Pro Plus software (4.1, Media Cybernetics Inc., Rockville, MD, USA) was used to trace two perpendicular lines to each other within each pore; these lines give the length between two points, so 30 measurements were made per sample to find the average diameter. Figure 2 shows the image used, illustrating the pore diameter of some of the measurements.

**Figure 2.** Measurement of the pore diameter using Image Pro Plus software.

The wall thickness measurements were performed using a scanning electron microscope (SEM) (Royal Philips Inc., Amsterdam, The Netherlands), as is shown in Figure 3 by a micrograph of the foam, indicating the measurements performed. To find the average wall thickness, 10 measurements per sample were performed.

**Figure 3.** Scanning electron microscope (SEM) micrograph of A-242 aluminium alloy foam where the wall thickness measurements are illustrated.

The microstructure of the A-242 aluminium alloy foams was observed using an XL30 ESEM scanning electron microscope (Royal Philips, Amsterdam, The Netherlands) equipped with a GENESIS 400 EDS (energy dispersion spectroscopy, Hi-Tech Instruments, Las Pinas, Philippines) microanalysis system to identify the intermetallic compounds present in the samples.

#### *2.3. Mechanical Characterization*

Finally, in order to evaluate the mechanical strength of the aluminium alloy foams under compression loads, uniaxial compression testing was carried out according to the ASTM E9 Standard procedure at room temperature. The tests were carried out using a Qtest Elite 100 model MTS electromechanical universal testing machine (MTS Inc., Berlin, Germany) with a capacity of 100 KN, equipped with TestWork software (Version 4, MTS Systems Corporation, Berlin, Germany). Cubic samples (25 × <sup>25</sup> × 25 mm3) were tested as shown in Figure 4. The compression tests were performed at cross-head rates of 3 mm/min. The force and the displacement were recorded during the compression tests. The engineering stress-strain data were determined through load-displacement measurements taking into account the initial dimensions of specimens.

**Figure 4.** Image of A-242 aluminium alloy foam, showing the closed porosity attained with 0.75% TiH2 at 948 K (A5).

#### **3. Results and Discussion**

#### *3.1. Foam Morphology*

Figure 4 shows an image of A-242 aluminium alloy foam obtained using the process described in this work: a structure of closed porosity having an average relative density of 0.12.

Figure 5 shows a photomicrograph of a A-242 aluminium alloy foam prepared with 0.75% TiH2 at 948 K (A5) using a stereographic microscope, where it can be observed that the pores of the sample are not connected to each other, thus A-242 aluminium alloy foams prepared using the technique described in this work presented a structure with closed porosity.

**Figure 5.** Stereographic photomicrography of A-242 aluminium alloy froam, obtained with 0.75% TiH2 at 948 K (A5).

The samples obtained have an average of 88.22% porosity, average pore size of 1.29 mm (0.50 mm standard deviation) and an average wall thickness of 114.67 μm (32.61 μm standard deviation). Table 5 presents the results obtained from the linear expansion, relative density, porosity, pore diameter, and wall thickness values for A-242 aluminium alloy foams obtained in the experiments indicated. We analyzed the effect of foaming temperature and TiH2 content on the properties of A-242 aluminium alloy foams.


**Table 5.** Linear expansion, relative density, porosity, pore diameter and wall thickness of the indicated A-242 aluminium alloy foams (*σ* = standard deviation).

As can observed from the values reported in Table 5, the foaming temperature does not greatly affect the linear expansion of the aluminium alloy foams obtained or the percentage of TiH2. Duarte et al. [40] report the effects of foaming temperature for 6061 alloy foams where they found that, when the foaming temperature is close to the solidus temperature, only a slight expansion occurs. If the foaming temperature is in the solid-liquid range, a greater expansion of the foam can be observed. However, increasing the temperature above the solid-liquid region reduces the viscosity of the alloy and promotes the production of more gas (H2) so that a greater expansion of the metal can be observed. Therefore, it is evident that the foaming process is sensitive to the foaming temperature chosen as well as the TiH2 content.

The relative density of A-242 aluminium alloy foams is highly sensitive to foaming temperature and TiH2 content. An increase in some of these parameters causes the relative density of the foam to decrease. The porosity of the foam is related to the relative density; therefore, this property is affected by both the foaming temperature and the content of TiH2.

The pore diameter and wall thickness of the pores is highly affected by the foaming temperature and the content of the foaming agent. An increase in foaming temperature produces a thinning of the pore walls effect of the coalescence and drainage phenomena of the foam (drained is a flow of molten metal from the walls into the pores edges (driven by surface tension) and through pore edges downwards driven by gravity). The same is true when a high content of H2 is released, giving rise to the phenomenon of coalescence (coalescence occurs whenever two pores merge to form a larger one) [40].

#### *3.2. Compression Behavior*

During the compression behavior of the foams, three characteristic zones must be evaluated: quasi-elastic, plateau, and densification [17,39].

The first area represents the quasi-elastic deformation behavior attained at smaller values of compression. A more complete analysis revealed that the deformation is partly reversible and a certain process of irreversible deformation of the foam structure occurs during the first load (depending on the density gradients, the structural composition, and microstructure of the foam). The second zone is that of constant stress or plateau, resulting in the abrupt and repeated failure of successive layers of pores (that in porous materials, the deformation takes place in low resistance regions, this being the one that presents the thinnest pore wall and the first contact zone during the compression test). If the foam is not perfect, then the stress in this zone shows ups and downs due to defects in the structure of the foam such as pore size distribution, low-density regions, very long pores, and walls of fractured pores. The third zone is that of densification; it begins when there are no longer enough walls of intact pores to withstand the load. The stress therefore increased quickly because the walls of the pores collided with each other, occupying the space left by the pores and causing the foam to densify, increasing its mechanical resistance [15,41]. Figure 6 shows the stress-strain curves of the A-242 aluminium alloy foams prepared to different processing parameters.

**Figure 6.** Effect of the percentage of TiH2 and temperature on the mechanical behavior of A-242 aluminium alloy foams, (**a**) 923 K, (**b**) 948 K, (**c**) 973 K.

In general, an increase in the content of the foaming agent added in the foaming process affects the internal structure of the aluminium alloy foams, i.e., by increasing pore size and decreasing resistance to deformation. The content of H2 retained in the aluminium alloy, which results from the thermal decomposition of the TiH2, causes the presence of large pores (greater than 2 mm) and low relative density values (less than 0.1). Therefore, the presence of intermetallic compounds such as Al3Fe and Al2CuMg causes brittleness during the compression test.

The samples foamed at 923 K presented very similar mechanical behaviors, with the sample foamed with 0.75 wt.% of TiH2 being the one that presented greater densification. The samples foamed at 948 K presented higher resistance compared with the samples obtained at 923 K. However, the foam with 1.00 wt.% TiH2 did not densify; this effect is due to its structure presenting large pores and low density. The aluminium alloy foams manufactured at 973 K presented a lower mechanical resistance, mainly for contents of 0.75 wt.% and 1.00 wt.% TiH2, the latter of which in general developed the lowest value of density. The exception to this was the aluminium alloy foam to which was added 0.50% TiH2, since this foam is the one that presented the highest value of resistance to deformation.

The most important characteristic that affects the mechanical properties of the foams is the relative density (relation between the density the foam and that of the solid). The metal foams have relative density values of less than 0.3 [14]. An increase in the density value is related to an increase in resistance.

The literature [14] establishes that the pore size of most metal foams lies in the range from 2 to 10 mm. Although the mechanical properties of the foams are sensitive to the wall thickness ratio (metal layer that separates one pore from another), most do not depend on the absolute pore size. The shape of the pores of the metal foams varies from equiaxial to ellipsoidal, and this has an important effect on mechanical behavior. In addition, the curvature of the walls and the chemical composition of the same affects the properties of the foam. Additives and foaming agents used to manufacture metal foams often result in unconventional alloys. An example of this is the use of Ca as a modifier of viscosity of the molten metal and TiH2 as a foaming agent, which introduces precipitates of Al, Ca, and Ti into the microstructure, weakening the walls of the pores [14]. The presence of calcium in the aluminium alloys forms intermetallic Al4Ca which presents a polyhedral morphology, being a precursor for the fracture and collapse of the foam during the compression test. In this case, the resistance of the aluminium alloy foam is expected to decrease in comparison with materials that do not present the formation of such intermetallic compounds.

Prieto et al. [42] presented a comparison of the stress-strain curves for aluminium foams obtained from the recycling of cans for beverages and pure aluminium, where the alloying elements such as Fe, Mn, and Mg form intermetallic compounds in the pore wall. The foam obtained from recycling presents a higher resistance than does the pure aluminium foam, an effect which is attributed to the presence of alloying elements such as Ca, Ti, Fe, and Mg which form intermetallic compounds such as Al4Ca, Al3Ti, and Al6(Fe, Mn).

Figure 7 shows a comparison of the stress-strain curves of foams obtained with pure aluminium, recycled aluminium cans, and the A-242 aluminium alloy obtained from the recycling of cans. The data of pure aluminium and of recycled aluminium were obtained from the work of Prieto et al. [42]. The pure aluminium and secondary aluminium foams have a relative density of 0.46, porosity of 82%, and pore diameter of 3 mm.

**Figure 7.** Stress–strain curves of a pure aluminium foam, an aluminium foam of an alloy prepared from beverage cans, and foams prepared with the A-242 aluminium alloy prepared in this work.

As can be observed in this figure, the foam of A-242 aluminium alloy prepared with 0.75 wt.% TiH2 at 948 K presents the highest mechanical strength (24.13 MPa) and short plateau zone, as compared to foams prepared with pure aluminium (13 MPa) and secondary aluminium prepared from beverage cans (21 MPa). For the preparation of A-242 aluminium alloy foams, a lesser Ca content (0.50 wt.%) is required to stabilize the aluminium alloy foam, as well as a lesser content of TiH2 to get low relative density values (0.13), a porosity of 82%, and a pore size of 1.5 mm. This could be an advantage for developing this kind of foam at an industrial level. In addition, due to the high Cu content in the A-242 aluminium alloy, the A-242 aluminium alloy foams can be thermally treated in order to improve their mechanical properties. The addition of TiH2 during foaming seems to also affect grain refinement during solidification; the TiAl3 particles formed improve the mechanical resistance of the aluminium alloy [43].

On the other hand, Ca particles seem to greatly affect the collapsing behavior of the foam during compression testing, as Ca forms new intermetallic compounds with Cu, weakening the foam wall and causing the fracture of the same before the application of the compression load [44]. Due to this fact, in this work we propose the use of a lower content of Ca to stabilize the aluminium alloy foam and avoid the formation of Cu-rich intermetallic compounds with calcium.

#### *3.3. Microstructural Analysis*

Figure 8 shows micrographs obtained via SEM where the pore walls of aluminium alloy foams with different contents of TiH2 are evident. The aluminium alloy foams depicted were obtained as follows: a) 0.50 wt.% TiH2, b) 0.75 wt.% TiH2, c) 1.00 wt.% TiH2 at 948 K. The intermetallic phases were identified using EDS in the SEM. The microstructure of the aluminium alloy foam cell wall is composed of aluminium dendrites, in addition to intermetallic compounds of Al2Cu (alternating lamellae of α-Al + θ-Al2Cu). Intermetallic compounds of the Al9FeNi and Al3Fe in their acicular forms and polymorphs of the Al2CuMg intermetallic compound are evident. The presence of Fe- and Mg-rich intermetallics comes from the raw material (beverage cans), in contrast to those rich in Ni and Cu, which come from the alloying elements added to adjust the chemical composition of the alloy. The presence of Ti-rich particles is a product of the decomposition reaction of the foaming agent (TiH2) during the process of frothing, where Al3Ti is formed by the reaction between aluminium and titanium. The cooling rate and the chemical composition of the aluminium foam are the main variables that determine the morphology, size and distribution of the different intermetallic compounds.

**Figure 8.** *Cont*.

**Figure 8.** Micrographs of A-242 aluminium alloy foams: (**a**) 0.50 wt.% TiH2; (**b**) 0.75 wt.% TiH2; and (**c**) 1.00 wt.% TiH2 at 948 K.

Figure 9 shows a micrograph obtained through SEM of a Ti-rich particle, surrounded by particles of the Al3Ti intermetallic compound, in an A-242 aluminium alloy foam prepared with 1.00 wt.% TiH2 at 948 K. The compound Al3Ti appears when the processing temperature is low, and the Al3Ti intermetallic gives the alloy greater mechanical strength [45]. The appearance of this intermetallic compound based on the equilibrium diagram is due to the preparation conditions used during the process (reaction time and foaming temperature). Figure 10 shows EDS patterns of the matrix, Ti-rich particle and Al3Ti intermetallic.

**Figure 9.** SEM micrograph of a Ti-rich particle, presumably the Al3Ti intermetallic compound.

**Figure 10.** Energy dispersion spectroscopy (EDS) patterns for the Ti-rich particle.

#### **4. Conclusions**

From the results obtained in this work, it is important to emphasize the possibility of using secondary aluminium for the preparation of aluminium alloy foams with characteristics similar to those obtained by the use of commercial-purity elements.

Aluminium alloy foams were prepared with closed porosity in the order of 88.22% and with a relative density of 0.12. The pore size was 1.29 mm, and the wall thickness was 114.67 μm.

Various intermetallic compounds were identified scattered on the pore walls, such as Al2Cu, Al9FeNi, Al3Fe, Al2CuMg, Ti, and Al3Ti. The aluminium alloy foams manufactured with 1.00 wt.% TiH2 showed higher concentrations of intermetallics.

With low TiH2 contents (0.50 wt.% and 0.75 wt.%), the samples presented high mechanical strength values; when analyzing the microstructure of this group of samples, it was observed that they present lower contents of intermetallic compounds distributed in the pore wall and the thickness of thinner wall.

The temperature of 948 K turned out to be the best temperature for foaming because the resulting aluminium alloy foams presented the highest values of compressive strength.

The aluminium alloy foams prepared from the A-242 aluminium alloy presented higher values of mechanical strength and lower relative density values when compared to foams prepared using commercial-purity elements.

**Author Contributions:** Investigation: N.M.T.R.; Supervision: J.T.T. and A.F.V.

**Funding:** This research was financed by the research stimulus program of CONACYT Mexico.

**Acknowledgments:** The authors thank the Consejo Nacional de Ciencia y Tecnología (CONACYT) and the Centro de Investigación y de Estudios Avanzados del Instituto Politécnico Nacional (CINVESTAV) Unidad Saltillo for their support in making the realization of this project possible.

**Conflicts of Interest:** The authors declare no conflicts of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*
