University of Groningen

Structural performance and failure analysis of aluminium foamsAmsterdam, Emiel

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite fromit. Please check the document version below.

Document VersionPublisher's PDF, also known as Version of record

Publication date:2008

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):Amsterdam, E. (2008). Structural performance and failure analysis of aluminium foams. s.n.

CopyrightOther than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of theauthor(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

The publication may also be distributed here under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license.More information can be found on the University of Groningen website: https://www.rug.nl/library/open-access/self-archiving-pure/taverne-amendment.

Take-down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediatelyand investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons thenumber of authors shown on this cover page is limited to 10 maximum.

Download date: 06-01-2022

1

Chapter 1

Introduction

1.1 Metal foam

Metal foam is a cellular solid, just like wood, coral, bone and bread, but with the cells made out of metal. Usually the metal is an aluminium alloy, but it can also be made of another metal, for example steel, nickel, titanium or gold. The relative density of a metal foam, which is the density of the metal foam divided by the density of the solid metal, should be less than 30% and it can be as low as a few percent [1]. Solid aluminium has a density of 2700 kg/m3 and the density of a metal foam made of aluminium is thus below 810 kg/m3 (=30%), making an aluminium foam of 30% relative density lighter than water (1000 kg/m3). Two types of metal foams can be distinguished: closed-cell and open-cell foam. In the closed-cell foam the cells are sealed off from its environment and the foam looks like bread (see Fig. 1.1). In the open-cell foam the cells are interconnected and the foam has the same appearance as a scouring pad (see Fig. 1.1). The connections in open-cell foam are called struts, which meet in nodes. For closed-cell foams the cells are closed by so-called faces, which meet in so-called cell edges, which in turn meet in nodes.

The first attempt to make metal foam started in 1943 with Benjamin Sosnick. He melted a mixture of aluminium and mercury in a high pressure chamber. When the pressure was released at the melting temperature of aluminium, the mercury evaporates and forms a foam [2,3,4,5]. In the 1950’s new techniques were developed, but it was still very hard to control the cell structure of the foam [6]. Improvements in the fabrication processes, as well as an increasing interest of industry in high performance and weight savings has revived the interest in functional materials such as metal foam. This has led to more research in the 1990’s. It has resulted in additional fabrication techniques, more applications and a better understanding of metal foam. This is nicely illustrated by the increase in the number of publications and patents (see Fig. 1.2) related to metal foam.

Chapter 1

2

Figure 1.1: Image of five cellular solids: Bread, an open- and closed-cell aluminium foam and a scouring pad. The last one on the background is the most commonly used cellular solid: wood.

Nowadays there are different fabrication processes for the open- and closed-cell foam. Most closed-cell aluminium foams are foamed in the liquid or the semi-solid state. Via the liquid state process foam is created by injecting air into molten aluminium, but just like water liquid aluminium does not foam. In water soap is added and in the aluminium solid particles, usually SiC, are added to keep the bubbles stable after rising to the surface. Afterwards the liquid metal foam is removed from the surface and cooled to maintain the foam structure and to create solid metal foam [5]. Due to this processing technique solid SiC particles are present in the aluminium matrix. Another technique for liquid state processing is by stirring blowing agents, usually TiH2 powder, into the molten aluminium. In the molten aluminium (T ~ 660°C) the TiH2 decomposes into Ti and H2 and the hydrogen gas forms bubbles. The bubbles are stabilized by adding Ca to the liquid aluminium, which increases the viscosity. After the foaming process the liquid

Introduction

3

Figure 1.2 (left) Number of publications vs. the publication year (from Isi Web of Knowledge). (right) Number of patents vs. the publication year (from www.wipo.int). Data has been obtained through a search on the 22nd of October 2007 using the following criteria: metal(lic) foam(s) or cellular metal(s) or alumin(i)um foam(s). foam is cooled to maintain the cellular structure [7]. As a result of this processing technique Ca and Ti is present in the aluminium matrix. In the semi-solid state processing route aluminium powder is mixed with TiH2 powder and compressed. The compacted mixture is placed inside a mould and heated to a temperature close to the melting point of aluminium, after which the TiH2 starts to decompose. When the right expansion is achieved the mould is cooled. For open-cell foam different fabrication techniques are available as well. Only a few are listed here. One way is by filling an open-cell polymer foam with a casting slurry. After curing the polymer foam is removed by heating and than the resulting mould is infiltrated with the molten metal. After cooling and removal of the mould material a metal foam is obtained which is a precise copy of the original polymer foam. Another way is by infiltrating compacted kitchen salt (NaCl), which has a melting temperature of 808°C, by molten metal. After cooling the salt is removed by dissolving it in water (For a complete description of the process procedures see Ref. [5] and Section 2.1). Fig. 1.3 shows the material properties in uniaxial tension and compression together with the definitions of the mechanical properties. Fig. 1.3a shows a typical stress-strain curve in tension, the straight line at the beginning of the curve is the linear elastic part, i.e. the stiffness of the foam, E*. The 0.2% offset yield stress, ∗

yσ ,

is obtained from the intersection of the curve with a line at 0.2% strain parallel to the linear elastic part of the curve. The peak stress, ∗

peakσ , is the highest stress

Chapter 1

4

(a)

(b)

Figure 1.3 Graph of a typical tensile (a) and compressive (b) stress-strain curves of a metal foam. E* is the stiffness of the foam, ∗

yσ the 0.2% yield stress, ∗peakσ the peak

stress, ∗peakε the peak strain, ∗

PCSσ the plastic collapse stress, ∗PCSε the strain at plastic

collapse and ∗Dε the densification strain.

obtained in the tensile test and the peak strain, ∗

peakε , is the strain at the peak stress.

Fig. 1.3b shows a typical stress-strain curve in compression. It is mainly characterized by the large plateau, which indicates that the foam can be compressed to a very large strain with a constant force. At a certain strain, ∗

Dε , the foam is compressed so much that all the cells are collapsed and the cell walls start to touch more and more leading to a densification of the foam. This makes it even harder to compress thereby increasing the stress. The stress at the small peak at the beginning

Introduction

5

of the plateau is denoted as the plastic collapse stress, ∗PCSσ . At that strain, ∗

PCSε , the

first cells start to collapse. The stiffness in tension and compression are about the same, just as the peak stress and the plastic collapse stress. Due to the cellular structure metal foam has properties which are interesting for applications. The following properties apply to open- and closed-cell aluminium foam:

• High stiffness to density ratio • High energy absorption during compression • High vibration absorption, both mechanical and acoustic • High temperature resistance, non-flammable • Electrical conductive, electromagnetic wave shielding • Thermal conductive • High interconnecting internal area (open-cell foam only) • Possibility of flow through the foam (open-cell foam only)

Additional properties:

• Non-toxic, corrosion and weather resistance • All kind of shapes (net-shape), low costs possible • Good machinability, Recyclable

The high stiffness to density ratio property of the foam is best utilized in bending structures as can be seen from the following example: The stiffness, S, of a flat panel is proportional to the stiffness of the panel material, E, and the thickness of the panel, h, to the power three: 3S Eh∝ . When the solid material of a panel is

replaced by a foam of the same material the stiffness of the panel material, i.e. the foam, changes by the relative density squared: 2

rE ρ∝∗ . For the same area and mass of the panel the thickness of the panel scales with inverse of the relative density: 1~ −

rfh ρ . This means that the stiffness of the panel scales with the inverse

of the relative density: 1323 −−∗∗ =∝∝ rrrfhES ρρρ . Since the relative density is lower

than unity the stiffness of the panel increases as the relative density decreases. Due to the cellular structure the metal foam can be compressed to very large strains. During compression the force is almost constant up the densification strain. Since the foam is made of a metal this force is very high compared to, for instance, a polymer foam. As a result of the high constant force over large strains, the energy

Chapter 1

6

absorption is very large in compression (see Fig. 1.3). The high vibration absorption is also related to the foam structure; the thin cell walls are capable of damping tiny vibrations and turn this energy into heat. For all the above stated individual properties, other materials or structures are more suitable. For example, waffle stiffened panels or honeycomb panels have a better stiffness to density ratio, but are more expensive and difficult to use for complex shapes. Additionally the mechanical properties of the honeycomb are very anisotropic. Ideally an application of metal foam utilizes a combination of properties (see Fig. 1.4). Attractive sectors for employing the high stiffness to density ratio and the vibration absorption are the transport- and machining industry. In the machining industry, metal foam can be used to replace structural machine parts that have to move back and forth with high precision. These parts made of foam can have the same stiffness with less mass and less vibrations. In the transport industry metal foam can be used to decrease the weight of car parts for the same stiffness, absorb vibration or sound and absorb impact energy during a crash.

Figure 1.4. Application categories for metal foam. A metal foam is best utilized in applications that employ a combination of properties. The primary colors, red, blue and yellow, stand for primary applications; in these categories other materials, such as honeycomb or waffle stiffened panels, are usually more suitable. The secondary colors stand for applications that use a combination of two metal foam properties. The white triangle in the middle represents a multifunctional application of metal foam.

Introduction

7

Examples of metal foam applications are: a support for a telescope working platform on a lorry, a cross slide of a high-speed milling machine and a crash-energy absorber (see Figs. 1.5 to 1.7). In general, the closed-cell foams are stronger than the open-cell foam and are therefore more used in structural applications. On the other hand open-cell foam has the advantage of an open structure, which can be used in applications such as: medical implants, filters, battery electrodes, fuel cells, heat sinks and degreasing filters in jet engines. Even tough the number of applications is increasing, the total number of applications is still rather limited. Several reasons can be listed: cheaper materials can be used for the same application, a possible additional step in the production process of the application and the lack of knowledge in industry regarding the mechanical properties, workability, fatigue behavior/long term use, etcetera. A lot of research in these areas has already been done in academia, but experience from working with real applications is only recently emerging. Therefore metal foam producing companies are now in a phase of optimizing and up-scaling the production process.

Figure 1.5 (left) New base of a lifting arm made from aluminium foam sandwich panels (AFS). The mass of the new base is 95 kg less than the steel variant and upon loading the deformation is 15% less. Manufacturer is Advanced Light-weight Materials GmbH in Saarbrücken, Germany. (right) A lorry with the new base, made from AFS panels, and lifting arm. The new base was designed to increase the vertical range of the platform from 20 m to 25 m while keeping the total weight of the vehicle below 3.5 ton, because otherwise the lorry would enter into a different drivers license category [8].

Chapter 1

8

Figure 1.6 Prototype of a BMW engine mounting bracket manufactured by LKR Ranshofen (Austria). From left to right: empty casting, entire composite part consisting of foam core and cast shell, section through composite part [8].

Figure 1.7 (left) Crash energy absorber for a tram built for the Combino vehicle system (right) Comino tram [8].

65 cm

Introduction

9

1.2 Scope of the thesis

As can be concluded from the applications of both the open- and closed-cell foams, high cyclic loads are usually present (except for the crash absorber). Due to the topology of the foam, compressive and tensile forces are always present under different loading conditions. This can lead to various failure mechanisms. Harte et al. claimed that the fatigue failure of sandwich beams in bending is due to microstructural damage in the foams, which is of tensile character [9]. Failure of a foam specimen in tension-tension fatigue is initiated by the following process: first microcracks appear which collapse to a large crack and the latter propagates through the sample leading to complete failure [10]. When and where these microcracks form, depends on the sample and the loading conditions. The effect of the sample can be divided in the effects of shape, topology and microstructure. Due to the production process the microstructure has been changed with respect to the bulk material. The effect of the altered microstructure on the crack initiation and propagation is not understood yet. The effect of the topology could for example be that during the first cycle of a tension-tension fatigue test the stress is concentrated in a certain area. At this high stress concentration a crack can already be introduced, which can than propagate during the rest of the test and cause early failure of the sample. This example can be derived from the fact that in monotonic uniaxial tensile tests damage is already introduced before the peak stress as mentioned in several papers [11,12,13,14,15]. However, at what stress the damage exactly initiates is not clear yet. Since tensile forces are always present within the foam structure and failure under bending is of tensile nature, our research is concentrated on the nucleation and propagation of damage under tensile loading: when do cracks form and how do they propagate? Therefore all subjects in this thesis are related to the understanding of the fracture behavior of metal foam in monotonic tension and fatigue. In this thesis damage is regarded as the necking or complete fracture of a strut or the rupture of a cell wall. There are several factors that could influence the mechanical response and the nucleation and evolution of damage in metal foam under tension, which can be subdivided in three categories:

Chapter 1

10

1. Sample properties • Relative density • Shape and size • Orientation of the long axis of the cells

2. Topology

• Cell size and shape, strut shape • Connectivity • Quality and density fluctuations

3. Microstructure

• Material properties • Precipitates

In the following section the importance and the role of these factors are further explained. Sample properties The relative density is an important parameter for the mechanical properties, such as the stiffness and the plateau stress [1]. The stiffness depends on the relative density by the following scaling relation:

21 rsECE ρ=∗ (1.1)

where Es is the stiffness of the base material, C1 a parameter that depends on the topology of the foam and ρr the relative density of the foam. For open-cell foams the plastic collapse stress is related to the relative density by the following scaling relation:

2/32 rysC ρσσ =∗ (1.2)

where σys is the yield stress of the base material and C2 is a parameter that depends again on the topology of the foam. How the relative density affects the damage evolution is unknown and one of the objectives of this thesis. As with tensile tests on solid test samples the shape and size of the sample has an influence on the mechanical properties of the foam. For a clean assessment of the mechanical properties the influence of the sample shape should be kept to a minimum. For tensile tests sharp corners, which act as stress concentrators, should be avoided and the length of the sample should be much larger than the width (or

Introduction

11

the squareroot of the cross-section). In general the size of the sample has to be seven times the cell size in each direction to obtain representative results [16,17]. Anisotropy of the mechanical properties is present in some types of metal foam. It is generally accepted that this is due to anisotropy in the cell shape, which in turn is caused by the particular production process [18,19,20,21,22]. Topology According to literature the cell size is not important to the mechanical properties such as the stiffness and the plateau stress. If this also holds for fatigue behavior is unknown. The shape of the cells and whether they are open or closed is important, because it determines the above mentioned constants C1 and C2, as well as the anisotropy in the mechanical properties. The connectivity, i.e. the number struts or cell edges that meet in a node, determines whether the deformation mechanism of the struts is bending or stretching dominated (see fig 1.8) [23,24]. Like a piece of paper, which is easy to bend and hard to stretch, it is also easier to bend a strut than to stretch a strut. Therefore a structure is much more rigid when the deformation mechanism of the struts is stretching dominated. Whether the deformation mechanism is bending or stretching dominated determines the power of the scaling law (Eqs. 1.1 and 1.2). For a low quality foam the number of defects, such as wiggled, non-uniform and missing cell walls, is very large [25,26]. This directly decreases the value of C1 and C2 in Eq. 1.1 and 1.2. Relative density fluctuations can occur over various length scales within the sample. This is inherent to the cellular architecture of the foam, but it depends very much on the production technique and is related to the quality of the foam. For example the presence of a very large cell, in metal foam samples with a non-uniform cell size distribution, with a similar cell wall thickness as the other cells locally decreases the relative density. Microstructure The effect of the microstructure can be separated into the effect of the base material properties and the precipitates. As already mentioned the mechanical properties of the foam depend on the mechanical properties of the base material [1]. These include the Young’ s modulus, yield stress, strain hardening exponent, UTS and failure strain. For example the presence of strain hardening in the base material totally changes the behavior of the metal foam as compared with ceramic foams, which are very brittle. These fail at the weakest link and failure is described by weakest link statistics. How the presence of plasticity/strain hardening exactly effects the fracture behavior is not understood. The mechanical properties of the

Chapter 1

12

(a)

(b)

Figure 1.8(a) Normalized stiffness vs. the relative density for different structures. (b) Normalized plastic collapse stress vs. the relative density for different structures. The figures are directly related to the scaling laws (Eqs. 1.1 and 1.2) and the slope in the figure is related to the power in the equations. The values of C1 and C2, which correspond to the quality and structure of a particular foam, determines the actual height for a certain relative density within the shaded area denoted as “foam”. Figures were obtained from Ref. [23].

Introduction

13

base material, such as the strain hardening exponent, can be altered by means of a precipitation heat treatment for some foams, without changing the sample or topology properties, to investigate the contribution of that property to the fracture behavior. The presence of brittle precipitates or inclusions due to the fabrication process can decrease the fracture strain of the struts without altering the material properties of the base material. In this thesis the effect of these factors are addressed whenever possible. All three subcategories/length scales were probed by using two closed-cell and two open-cell foams, various samples configurations, set-ups and measuring techniques. In the prologue of each chapter the factors, which are probed in that chapter, are mentioned as well as the correlation to the other chapters. The outline of the thesis is as follows: in Chapter 2 the production processes and standard experimental tools are described. In Chapter 3 the influence of the microstructure on the crack initiation of Metcomb closed-cell foam in tension is investigated. In Chapter 4 the influence of the microstructure on the fracture behavior of Alporas closed-cell foam in tension and fatigue is described. In Chapters 5 to 8 the effect of the relative density, the orientation of the long axis of the cell, the strut shape, the heat treatment and the presence of brittle precipitates is investigated for Duocel open-cell foam. In Chapter 9 the findings on Duocel open-cell foam are compared to the results obtained for an open-cell foam with the same density range, but with a different production process and topology. The thesis concludes with an outlook, list of publications and summary.

References

1. Gibson LJ, Ashby MF. Cellular Solids, Structure & Properties (Pergamon Press, 1st ed. 1988). 2. B. Sosnick. US Patent 2434775 (1948). 3. Davies GJ, Zhen S. J Mat Sci 1983;18;1899. 4. Banhart J, Weaire D. Phys Today 2002;7;37. 5. Banhart J. Progress Mater Sci 2001;46;559. 6. Elliott JC. US patent 4099961 (1956). 7. Miyoshi T, Itoh M, Akiyama S, Kitahara A. Adv Eng Mat 2000;2;179. 8. Banhart J. Int J Vehicle Design 2005;37;114. 9. Harte A-M, Fleck NA, Ashby MF. Int J Fatigue 2001;23;499-507. 10. Amsterdam E, De Hosson JTM, Onck PR. Acta Mater 2006;54;4465.

Chapter 1

14

11. Motz C, Pippan R. Acta Mater 2002;50;2013. 12. San Marchi C, Despois J-F, Mortensen A. Acta Mater 2004;52;2895. 13. Dillard T, N’Guyen F, Maire E, Salvo L, Forest S, Bienvenu Y, Bartout JD, Croset M, Dendievel R, Cloetens P. Phil Mag 2005;85;2147. 14. Amsterdam E, Onck PR, De Hosson JTM. J Mater Sci 2005;40;5813. 15. Amsterdam E, Vries JHB de, De Hosson JTM, Onck PR. Acta Mater 2008;56;609. 16. Onck PR, Andrews EW, Gibson LJ. Int J Mech Sci 2001;43: 681. 17. Andrews EW, Gioux G, Onck PR, Gibson LJ. Int J Mech Sci 2001;43:701. 18. McCullough KYG, Fleck NA, Ashby MF. Acta Mater 1999;47;2323. 19. Andrews E, Sanders W, Gibson LJ. Mater Sci Eng A 1999;270;113. 20. Olurin OB, Fleck NA, Ashby MF. Mater Sci Eng A 2000;291;136. 21. Badiche X, Forest S, Guibert T, Bienvenu Y, Bartout J-D, Ienny P, Croset M, Bernet H. Mater Sci Eng A 2000;289;276. 22. Nieh TG, Higashi K, Wadsworth J. Mater Sci Eng A 2000;283;105. 23. Deshpande VS, Ashby MF, Fleck NA. Acta Mater 2001;49;1035. 24. Ashby MF. Phil Trans R Soc A 2006;365;15. 25. Evans AG, Hutchinson JW, Ashby MF. Progress Mater Sci 1999;43;171. 26. McCullough KYG, Fleck NA, Ashby MF. Acta Mater 1999;47;2323.