**Porous Metals and Metal Foams Made from Powders**

Andrew Kennedy *Manufacturing Division, University of Nottingham, Nottingham, UK* 

### **1. Introduction**

30 Powder Metallurgy

Zhigao, X. and Xiaofei, L. 1991. A Research for the Friction and Wear Properties of a Metal-

Science Publisher. pp. 611-615.

fiber-reinforced Composite Material. *In Mechanical Properties Materials Design. International Symposia Proceedings* 1991. Boqun Wu (ed.). Amsterdam: Elsevier

> Porous materials are found in natural structures such as wood, bone, coral, cork and sponge and are synonymous with strong and lightweight structures. It is not surprising that manmade porous materials have followed and those made from polymers and ceramics have been widely exploited. The commercialisation of porous metals has lagged somewhat behind, but has received a boost following a surge in worldwide research and development in the early 1990's.

> The unique combination of physical and mechanical properties offered by porous metals, combinations that cannot be obtained with dense metals, or either dense or porous polymers and ceramics, makes them attractive materials for exploitation. Interest mainly focuses on exploiting their ability to be incorporated into strong, stiff lightweight structures, particularly those involving Al foams as the "filling" in sandwich panels, their ability to absorb energy, vibration and sound and their resilience at high temperature coupled with good thermal conductivity.

> The applications for porous metals (metals having a large volume of porosity, typically 75- 95%) and metal foams (metals with pores deliberately integrated into their structure through a foaming process) depend on their structure. Closed cell foams, which have gas-filled pores separated from each other by metal cell walls, have good strength and are mainly used for structural applications. Open cell foams, which contain a continuous network of metallic struts and the enclosed pores in each strut frame are connected (in most cases these materials are actually porous or cellular metals), are weaker and are mainly used in functional applications where the continuous nature of the porosity is exploited. Examples of structures for both these types of porous metals are shown in Figure 1.

> Table 1 summarises the potential uses for porous metals and metal foams, highlighting the relevant attributes that make them suitable for that particular application (Ashby et al., 2000). Specific examples of current applications for porous metals and metal foams can be found in (Banhart, 2001), but it should be remarked that this is a dynamic area and new applications and components are continually being developed.

> This short review describes the main powder metallurgy-based manufacturing routes, highlighting the key aspects of the relevant technology involved and the types of foam structures that result.

Porous Metals and Metal Foams Made from Powders 33

into and sintered in a die. With packing densities broadly in the range of 40-60%, but affected by particle shape, size and vibration, the porosities in these structures are well below those for most porous metals. The simplicity of the process means that porosity can be included in a wide range of metals, limited only by the ability to sinter the metal in an appropriate die. The process is most commonly used to sinter bronze powders to make bearings, an example of which is shown in Figure 2, but porous structures from titanium,

Fig. 2. Porous bronze made by pressureless sintering of approximately 100 m diameter

In the case of Al alloys, where sintering is made difficult by the surface oxide layer covering the powder particles, milling of the powder with sintering aids such as Sn and Mg is required. The limitations imposed by the need to sinter in a die, principally on the size of the product and the productivity, can be mitigated by performing die (or other) compaction processes to increase the green strength of the compact so that it is sufficient to perform containerless sintering. The inevitable consolidation that is involved will, of course, decrease

In an effort to increase the porosity in these parts, powders have been replaced by metal fibres, made by processes such as melt spinning, which with higher aspect ratios, exhibit lower packing densities, making them suitable for a wider range of applications (Anderson and Stephani, 1999). A further development of this, leading to much higher porosities, is the sintering of hollow metal spheres (which themselves are made via a powder route – to be described later). Spheres with diameters ranging from 1.5 to 10 mm, with wall thicknesses from 20 to 500 m can be arranged to produce both open and closed pore structures. Open structures can be obtained in the same way as for powders and compaction can be used to deform the spheres to polyhedral bodies, reducing the degree of open porosity. True closed pore structures can be obtained by filling the interstices between the spheres with metal powder followed by a sintering treatment. Porosity is contained both within and between the hollow spheres and porosities in the range of 80-97% are reported (Anderson et al., 2000). Examples of an open cell sintered hollow sphere structure and the spheres in cross

section, showing their thin walls, are presented in Figure 3.

superalloys and stainless steel have also been made in this way.

powders [Eisenmann, 1998].

the already low porosity.

Fig. 1. Micrographs of (left) a closed cell foam and (right) an open cell foam, or more correctly, a cellular metal (Zhou, 2006).


Table 1. Potential application areas for porous metals and metal foams adapted from (Ashby et al., 2000).

#### **2. Processing methods for metal foams**

There are many different ways to produce porous metals and metallic foams and these methods are usually classified into four different types of production, using liquid metals, powdered metals, metal vapour or metal ions. The use of powdered metals as the starting material for foam production offers the same types of advantages (and often the same limitations) as conventional powder metallurgical processes. If a particular metal or alloy can be pressed and sintered, there is a high likelihood that it can be made into a porous metal or metal foam.

#### **2.1 Porous metals produced by powder sintering**

#### **2.1.1 Pressureless sintering**

Loose pack, pressureless or gravity sintered metal powders were the first form of porous metals and are still widely used as filters and as self-lubricating bearings. The porosity in these components is simply derived from the incomplete space filling of powders poured

Fig. 1. Micrographs of (left) a closed cell foam and (right) an open cell foam, or more

Lightweight structures Excellent stiffness-to-weight ratio when loaded in bending Mechanical damping Damping capacity is larger than solid metals by up to 10x Vibration control Foamed panels have higher natural flexural vibration

Acoustic absorption Open cell metal foams have sound-absorbing capacity

Heat exchangers Open-cell foams have large accessible surface area and high

Filters Open-cell foams for high-temperature gas and fluid filtration Table 1. Potential application areas for porous metals and metal foams adapted from (Ashby

There are many different ways to produce porous metals and metallic foams and these methods are usually classified into four different types of production, using liquid metals, powdered metals, metal vapour or metal ions. The use of powdered metals as the starting material for foam production offers the same types of advantages (and often the same limitations) as conventional powder metallurgical processes. If a particular metal or alloy can be pressed and sintered, there is a high likelihood that it can be made into a porous

Loose pack, pressureless or gravity sintered metal powders were the first form of porous metals and are still widely used as filters and as self-lubricating bearings. The porosity in these components is simply derived from the incomplete space filling of powders poured

pressure

Biocompatible inserts Cellular texture stimulates cell growth

frequencies than solid sheet of the same mass per unit area

cell-wall conduction giving exceptional heat transfer ability

Exceptional ability to absorb energy at almost constant

correctly, a cellular metal (Zhou, 2006).

Energy absorbers /

packaging

et al., 2000).

metal or metal foam.

**2.1.1 Pressureless sintering** 

Application Relevant Attributes

**2. Processing methods for metal foams** 

**2.1 Porous metals produced by powder sintering** 

into and sintered in a die. With packing densities broadly in the range of 40-60%, but affected by particle shape, size and vibration, the porosities in these structures are well below those for most porous metals. The simplicity of the process means that porosity can be included in a wide range of metals, limited only by the ability to sinter the metal in an appropriate die. The process is most commonly used to sinter bronze powders to make bearings, an example of which is shown in Figure 2, but porous structures from titanium, superalloys and stainless steel have also been made in this way.

In the case of Al alloys, where sintering is made difficult by the surface oxide layer covering the powder particles, milling of the powder with sintering aids such as Sn and Mg is required. The limitations imposed by the need to sinter in a die, principally on the size of the product and the productivity, can be mitigated by performing die (or other) compaction processes to increase the green strength of the compact so that it is sufficient to perform containerless sintering. The inevitable consolidation that is involved will, of course, decrease the already low porosity.

In an effort to increase the porosity in these parts, powders have been replaced by metal fibres, made by processes such as melt spinning, which with higher aspect ratios, exhibit lower packing densities, making them suitable for a wider range of applications (Anderson and Stephani, 1999). A further development of this, leading to much higher porosities, is the sintering of hollow metal spheres (which themselves are made via a powder route – to be described later). Spheres with diameters ranging from 1.5 to 10 mm, with wall thicknesses from 20 to 500 m can be arranged to produce both open and closed pore structures. Open structures can be obtained in the same way as for powders and compaction can be used to deform the spheres to polyhedral bodies, reducing the degree of open porosity. True closed pore structures can be obtained by filling the interstices between the spheres with metal powder followed by a sintering treatment. Porosity is contained both within and between the hollow spheres and porosities in the range of 80-97% are reported (Anderson et al., 2000). Examples of an open cell sintered hollow sphere structure and the spheres in cross section, showing their thin walls, are presented in Figure 3.

Porous Metals and Metal Foams Made from Powders 35

Fig. 4. A schematic (left) of the processing steps used to manufacture titanium alloy foam sandwich panels by gas entrapment (Ashby et al., 2000) and (right) the morphology of a

reactions, initiated either by local or global heating of compacted powder mixtures to the reaction ignition temperature, lead to vapourisation of hydrated oxides on the powder surfaces and the release of gases dissolved in the powder. The reacting powder mixture heats up rapidly to form a liquid containing (mostly hydrogen) gas bubbles and when the reaction is complete, cools rapidly, entrapping the gas to form a foam. A schematic of this

Gas formation and foam expansion can be augmented by the addition of vapour forming phases such as carbon (which burns in air to produce CO) or foaming agents which react together to increase the reaction temperature and produce fine particles that stabilise the foam. As foaming takes place in the liquid state, stabilisation of the bubbles is needed to avoid rapid collapse of the foam structure. Figure 5 shows how a reactive Ti+B4C foaming agent increases the porosity in a Ni-Al powder mixture from 30 to 90% by 5% addition. It can be seen that the pores are irregular in shape, as is the shape of the expanded foam. Although the process is relatively simple, the production of foams is limited to combinations of materials that react exothermically, some metal-metal systems but typically metals and carbon or carbides, or metals and oxides. These limitations, coupled with the difficulty controlling the expansion process and defining the shape of the expanded foam, mean that

TiAl6V4 sandwich structure (Banhart, 2001).

process is shown in Figure 5 (Kanetake and Kobashi, 2006).

few, if any, commercial foam products are made this way.

Fig. 3. Steel hollow spheres (left) sintered to form an open cell foam and (right) sectioned.

The advantage of porous structures made from sintered powders or hollow spheres is that there is good control of the volume fraction and to a lesser extent, the geometry and size of the pores, leading to reproducible structures and properties. Sintered metal powder and fibre structures are mainly suited to applications based on filtration, catalysis or heat exchange. The low density structures that can be made from sintered hollow spheres can be used for lightweight structural parts and for energy and sound absorption.

#### **2.1.2 Gas entrapment**

Internal porosity can also be developed in metal structures by a gas expansion (or foaming) process based on hot isostatic pressing (HIPing) (Martin and Lederich, 1992). Initially, following the standard method for HIPing of metal powders, a gas-tight metal can is filled with powder and evacuated. In a deviation from common practise, the can is then filled with argon gas at pressures between 3 and 5 bar before being sealed, isostatically pressed at high temperature and then worked to form a shaped product, normally a sheet. Porosity is generated by annealing the part. When holding at elevated temperature, the pressurised argon gas present within small pores in the structure causes the material to expand (foam) by creep. As HIPing is an effective method for sintering and the can material can be made from the same material as the powder, this process could be used for many different metal powders. The use of a can means that sandwich-type structures, consisting of a lightweight foam core and thin, solid face sheets are produced. These types of structures are ideal for lightweight construction.

Porous bodies with typically 20–40% of isolated porosity are obtained and theoretical considerations show that no more than 50% porosity can be expected (Elzey and Wadley, 2001). Figure 4 shows a schematic representation of the process, which has been used to make porous titanium sandwich structures for the aerospace industry, without the need for complex joining methods. Disadvantages include low porosity and irregular-shaped pores.

#### **2.1.3 Reactive processing**

In contrast to the gas entrapment method, porosity is evolved much more rapidly when foaming occurs in highly reactive multi-component powder systems such as those which undergo self-propagating high temperature synthesis (SHS). The highly exothermic

Fig. 3. Steel hollow spheres (left) sintered to form an open cell foam and (right) sectioned.

used for lightweight structural parts and for energy and sound absorption.

**2.1.2 Gas entrapment** 

lightweight construction.

**2.1.3 Reactive processing** 

The advantage of porous structures made from sintered powders or hollow spheres is that there is good control of the volume fraction and to a lesser extent, the geometry and size of the pores, leading to reproducible structures and properties. Sintered metal powder and fibre structures are mainly suited to applications based on filtration, catalysis or heat exchange. The low density structures that can be made from sintered hollow spheres can be

**2mm**

Internal porosity can also be developed in metal structures by a gas expansion (or foaming) process based on hot isostatic pressing (HIPing) (Martin and Lederich, 1992). Initially, following the standard method for HIPing of metal powders, a gas-tight metal can is filled with powder and evacuated. In a deviation from common practise, the can is then filled with argon gas at pressures between 3 and 5 bar before being sealed, isostatically pressed at high temperature and then worked to form a shaped product, normally a sheet. Porosity is generated by annealing the part. When holding at elevated temperature, the pressurised argon gas present within small pores in the structure causes the material to expand (foam) by creep. As HIPing is an effective method for sintering and the can material can be made from the same material as the powder, this process could be used for many different metal powders. The use of a can means that sandwich-type structures, consisting of a lightweight foam core and thin, solid face sheets are produced. These types of structures are ideal for

Porous bodies with typically 20–40% of isolated porosity are obtained and theoretical considerations show that no more than 50% porosity can be expected (Elzey and Wadley, 2001). Figure 4 shows a schematic representation of the process, which has been used to make porous titanium sandwich structures for the aerospace industry, without the need for complex joining methods. Disadvantages include low porosity and irregular-shaped pores.

In contrast to the gas entrapment method, porosity is evolved much more rapidly when foaming occurs in highly reactive multi-component powder systems such as those which undergo self-propagating high temperature synthesis (SHS). The highly exothermic

Fig. 4. A schematic (left) of the processing steps used to manufacture titanium alloy foam sandwich panels by gas entrapment (Ashby et al., 2000) and (right) the morphology of a TiAl6V4 sandwich structure (Banhart, 2001).

reactions, initiated either by local or global heating of compacted powder mixtures to the reaction ignition temperature, lead to vapourisation of hydrated oxides on the powder surfaces and the release of gases dissolved in the powder. The reacting powder mixture heats up rapidly to form a liquid containing (mostly hydrogen) gas bubbles and when the reaction is complete, cools rapidly, entrapping the gas to form a foam. A schematic of this process is shown in Figure 5 (Kanetake and Kobashi, 2006).

Gas formation and foam expansion can be augmented by the addition of vapour forming phases such as carbon (which burns in air to produce CO) or foaming agents which react together to increase the reaction temperature and produce fine particles that stabilise the foam. As foaming takes place in the liquid state, stabilisation of the bubbles is needed to avoid rapid collapse of the foam structure. Figure 5 shows how a reactive Ti+B4C foaming agent increases the porosity in a Ni-Al powder mixture from 30 to 90% by 5% addition. It can be seen that the pores are irregular in shape, as is the shape of the expanded foam. Although the process is relatively simple, the production of foams is limited to combinations of materials that react exothermically, some metal-metal systems but typically metals and carbon or carbides, or metals and oxides. These limitations, coupled with the difficulty controlling the expansion process and defining the shape of the expanded foam, mean that few, if any, commercial foam products are made this way.

Porous Metals and Metal Foams Made from Powders 37

Fig. 6. Production of an open cell foam by sintering a mixture of metal powder and a

suitable solvent in which the metal matrix should be unaffected, most commonly water. In the case of Al alloys, NaCl is the favoured space holder as it melts at approximately 800°C and can be dissolved fairly rapidly in warm water. For higher melting point metals, NaAlO2,

In addition to the decomposition or melting temperature and general ease of processing, including solubility in the solvent, space fillers are also selected based on their inertness and lack of solubility in the matrix, as well as the availability in the size and shape desired (typically in the range of hundreds of m to a few mm and often spherical space fillers are preferred). As salts (such as NaCl) are brittle materials and, particularly when ground, are angular in shape, melting and granulation methods have been used to produce them in spherical form (Goodall and Mortensen, 2007). If these granules are porous, this has the additional benefit of more rapid removal from the sintered part due to the granule being able to disintegrate as well as dissolve. Figure 7 shows porous beads made by the controlled agglomeration of fine salt particles and the structure of the resulting porous metal (Jinnapat and Kennedy, 2010). The cell structure is clearly interconnected via windows between neighbouring pores, the number and size of which are dependent upon the co-ordination

Processing difficulties can arise during mixing the metal powder with the often much larger space holders. Segregation results in defects in the final product, usually in the form of incomplete struts and cells. Liquid binders are used to promote homogeneous mixing, often by ensuring that the space filler is coated with the metal powder. An alternative approach is to vibrate the finer metal powders into the interstices within a packed bed of the larger space holder particles. The metal powder – space holder mixtures are then compressed to form a compact with sufficient strength to be sintered free-standing and, in the case of Al to disrupt surface oxide films so that sintering can take place. Compaction processes can include all of

those used in standard metallurgical practices, including metal injection moulding.

removable agent (Ashby et al., 2000).

with a melting point of 1800°C, is used.

number and contact area for the packed spheres.

Fig. 5. A schematic (top) of the reactive powder process used to make metal foams and (bottom) cross-sections of NiAl3 foams containing different addition levels of a Ti + B4C foaming agent mixture, after (Kanetake and Kobashi, 2006).

#### **2.1.4 The addition of space-holding fillers**

Many of the processes already mentioned have shortcomings in either the rather limited levels of porosity that are achievable, or in the formation of pores that are highly irregular with a wide distribution in sizes. A simple development of standard PM practices, but incorporating a volume of sacrificial space fillers, offers a solution to both these problems.

Porous metals are produced by mixing and compacting metal powders with a space holder which is later removed either during or after sintering, by dissolution or thermal degradation, to leave porosity (Zhao and Sun, 2001). A schematic of this process is shown in Figure 6. This simple method has the advantage that the morphologies of the pores and their size are determined by the characteristics of the space holder particles and the foam porosity can be easily controlled by varying the metal/space holder volume ratio. Addition levels of the space holder are typically between and 50% and 85%. These are sufficiently high that they are interconnected and hence can be removed easily. Above 85% the structure of the struts is unlikely to be continuous and below 50%, residual space filler will be enclosed within the structure, making removal very difficult.

Commonly, space fillers take the form of polymer granules and water soluble salts, but can also be metal powders and ceramic or polymer hollow spheres (if the hollow spheres aren't removed then these materials are, strictly speaking, syntactic foams). Two distinct approaches may be taken. In the first, the space filler can either thermally decompose, sublime or evaporate below the sintering temperature of the metal matrix. This requires the resulting skeletal metal structure to have good strength, which is affected during the compaction stage. Polymer powders (for example PMMA), carbamide granules or Mg grains are commonly used. In the second approach, the space filler has a higher melting point than the sintering temperature and is removed, after sintering, by dissolution in a

Fig. 5. A schematic (top) of the reactive powder process used to make metal foams and (bottom) cross-sections of NiAl3 foams containing different addition levels of a Ti + B4C

Many of the processes already mentioned have shortcomings in either the rather limited levels of porosity that are achievable, or in the formation of pores that are highly irregular with a wide distribution in sizes. A simple development of standard PM practices, but incorporating a volume of sacrificial space fillers, offers a solution to both these problems. Porous metals are produced by mixing and compacting metal powders with a space holder which is later removed either during or after sintering, by dissolution or thermal degradation, to leave porosity (Zhao and Sun, 2001). A schematic of this process is shown in Figure 6. This simple method has the advantage that the morphologies of the pores and their size are determined by the characteristics of the space holder particles and the foam porosity can be easily controlled by varying the metal/space holder volume ratio. Addition levels of the space holder are typically between and 50% and 85%. These are sufficiently high that they are interconnected and hence can be removed easily. Above 85% the structure of the struts is unlikely to be continuous and below 50%, residual space filler will be enclosed

Commonly, space fillers take the form of polymer granules and water soluble salts, but can also be metal powders and ceramic or polymer hollow spheres (if the hollow spheres aren't removed then these materials are, strictly speaking, syntactic foams). Two distinct approaches may be taken. In the first, the space filler can either thermally decompose, sublime or evaporate below the sintering temperature of the metal matrix. This requires the resulting skeletal metal structure to have good strength, which is affected during the compaction stage. Polymer powders (for example PMMA), carbamide granules or Mg grains are commonly used. In the second approach, the space filler has a higher melting point than the sintering temperature and is removed, after sintering, by dissolution in a

foaming agent mixture, after (Kanetake and Kobashi, 2006).

**2.1.4 The addition of space-holding fillers** 

within the structure, making removal very difficult.

Fig. 6. Production of an open cell foam by sintering a mixture of metal powder and a removable agent (Ashby et al., 2000).

suitable solvent in which the metal matrix should be unaffected, most commonly water. In the case of Al alloys, NaCl is the favoured space holder as it melts at approximately 800°C and can be dissolved fairly rapidly in warm water. For higher melting point metals, NaAlO2, with a melting point of 1800°C, is used.

In addition to the decomposition or melting temperature and general ease of processing, including solubility in the solvent, space fillers are also selected based on their inertness and lack of solubility in the matrix, as well as the availability in the size and shape desired (typically in the range of hundreds of m to a few mm and often spherical space fillers are preferred). As salts (such as NaCl) are brittle materials and, particularly when ground, are angular in shape, melting and granulation methods have been used to produce them in spherical form (Goodall and Mortensen, 2007). If these granules are porous, this has the additional benefit of more rapid removal from the sintered part due to the granule being able to disintegrate as well as dissolve. Figure 7 shows porous beads made by the controlled agglomeration of fine salt particles and the structure of the resulting porous metal (Jinnapat and Kennedy, 2010). The cell structure is clearly interconnected via windows between neighbouring pores, the number and size of which are dependent upon the co-ordination number and contact area for the packed spheres.

Processing difficulties can arise during mixing the metal powder with the often much larger space holders. Segregation results in defects in the final product, usually in the form of incomplete struts and cells. Liquid binders are used to promote homogeneous mixing, often by ensuring that the space filler is coated with the metal powder. An alternative approach is to vibrate the finer metal powders into the interstices within a packed bed of the larger space holder particles. The metal powder – space holder mixtures are then compressed to form a compact with sufficient strength to be sintered free-standing and, in the case of Al to disrupt surface oxide films so that sintering can take place. Compaction processes can include all of those used in standard metallurgical practices, including metal injection moulding.

Porous Metals and Metal Foams Made from Powders 39

Fig. 8. Cellular structures made from 316L stainless steel by direct typing (Andersen et al.,

Slurry processing of ceramics has been used for many years to produce bulk products, coatings, films and foams. The wealth of scientific research and resulting literature pertaining to this area is not mirrored for metal powders slurries, the need for processing in this way is perhaps not so great given the many alternative forming methods for metals. The production of stable (non-agglomerated, non-sedimenting) slurries containing metal powders provides significant challenges. Metal powders are larger in size than those used for ceramic processing and have a higher density, making them more difficult to keep suspended. Metal powder slurries, like their ceramic counterparts, are usually based on aqueous systems to which a suspending agent (to increase the viscosity) and dispersants (to prevent particle flocculation) are added (Kennedy and Lin, 2011). Other additions may be

The advantages of using metal slurries as the basis for producing foams is that constraints imposed by poor compressibility are removed, shaping can be performed by gelling in simple rubber moulds and gas bubbles can be introduced into the system by whisking or through the addition of gas-forming agents (although this means that the foam has to be stabilised). Although not well established, there are likely to be limitations to the materials that can be used based on interactions between the metal and the solvent (corrosion or

Porous metals can be produced by a replication method using an open cell polyurethane (PU) foam or sponge as a template (Quadbeck et al., 2007). In this process the polymer foam is first coated with a metal powder slurry, usually performed by immersion or by spraying. Excess slurry is removed by squeezing the foam, often by passing it through rollers. Without this, the cells may become partially closed due to the formation of liquid films bridging the cell struts. After coating and drying, the template is removed by thermal degradation and the resulting, fragile metal skeleton is further heated to sinter the metal powder particles,

needed to lower the surface tension, change the pH or facilitate gelation.

2004).

**2.2 Metal powder slurry processing** 

reaction) and the ability to form a stable slurry.

**2.2.2 Slurry coating of polyurethane foams** 

forming a rigid cellular metal structure.

**2.2.1 Metal powders slurries** 

Fig. 7. Images showing (left) porous spherical salt beads and (right) the microstructure of the resulting cellular stainless steel (Jinnapat and Kennedy, 2010).

Despite a few drawbacks, for example if space fillers such as NaCl are not completely removed this can lead to corrosion and the size of parts produced is rather limited, in part due to slow dissolution of the space fillers which might take days even for small parts, this method is a favoured route for the manufacture of porous metals from a wide range of materials. It is particularly suited to those metals with high melting points and is a common route for the production of porous biomedical devices made from Ti or NiTi.

#### **2.1.5 Additive manufacturing**

Porous metal structures can be built, layer upon layer, using processes such as selective laser sintering (SLS), direct metal typing or 3D direct metal printing. 3D parts are constructed by stacking these layers, the geometry of which is defined by a CAD model.

Direct metal laser sintering uses a high power laser to sinter metal powders on the surface of a powder bed. After each cross-section is scanned, the powder bed is lowered by one layer thickness, a new layer of material is applied on top, and the process is repeated until the part is completed. Two-component powders can be used, comprising metals coated with a polymer, where the laser only melts the coating. Products made in this way, however, still need a secondary sintering step to produce sufficiently robust parts. Steel and titanium foams can be made by both these methods, although with polymeric binders, contamination of Ti with carbon occurs and direct metal laser sintering is, therefore, preferred.

Direct metal printing works in a similar way, spraying a polymeric binder, which is dispensed through a print head, over a powder bed. 3D metal typing or screen printing uses a metal powder mixed with a binder which is then spread over patterned masks and cured layer by layer (Andersen et al., 2004). For both these processes, subsequent polymer-removal and sintering steps are required. Figure 8 shows cellular structures made from 316L stainless steel by direct typing. Among the benefits of these processes are high levels of design flexibility, including small, precise cellular structures with complex internal geometries.

Fig. 7. Images showing (left) porous spherical salt beads and (right) the microstructure of the

Despite a few drawbacks, for example if space fillers such as NaCl are not completely removed this can lead to corrosion and the size of parts produced is rather limited, in part due to slow dissolution of the space fillers which might take days even for small parts, this method is a favoured route for the manufacture of porous metals from a wide range of materials. It is particularly suited to those metals with high melting points and is a common

Porous metal structures can be built, layer upon layer, using processes such as selective laser sintering (SLS), direct metal typing or 3D direct metal printing. 3D parts are constructed by

Direct metal laser sintering uses a high power laser to sinter metal powders on the surface of a powder bed. After each cross-section is scanned, the powder bed is lowered by one layer thickness, a new layer of material is applied on top, and the process is repeated until the part is completed. Two-component powders can be used, comprising metals coated with a polymer, where the laser only melts the coating. Products made in this way, however, still need a secondary sintering step to produce sufficiently robust parts. Steel and titanium foams can be made by both these methods, although with polymeric binders, contamination

Direct metal printing works in a similar way, spraying a polymeric binder, which is dispensed through a print head, over a powder bed. 3D metal typing or screen printing uses a metal powder mixed with a binder which is then spread over patterned masks and cured layer by layer (Andersen et al., 2004). For both these processes, subsequent polymer-removal and sintering steps are required. Figure 8 shows cellular structures made from 316L stainless steel by direct typing. Among the benefits of these processes are high levels of design flexibility, including small, precise cellular structures with complex internal geometries.

route for the production of porous biomedical devices made from Ti or NiTi.

stacking these layers, the geometry of which is defined by a CAD model.

of Ti with carbon occurs and direct metal laser sintering is, therefore, preferred.

resulting cellular stainless steel (Jinnapat and Kennedy, 2010).

**2.1.5 Additive manufacturing** 

Fig. 8. Cellular structures made from 316L stainless steel by direct typing (Andersen et al., 2004).

#### **2.2 Metal powder slurry processing**

#### **2.2.1 Metal powders slurries**

Slurry processing of ceramics has been used for many years to produce bulk products, coatings, films and foams. The wealth of scientific research and resulting literature pertaining to this area is not mirrored for metal powders slurries, the need for processing in this way is perhaps not so great given the many alternative forming methods for metals. The production of stable (non-agglomerated, non-sedimenting) slurries containing metal powders provides significant challenges. Metal powders are larger in size than those used for ceramic processing and have a higher density, making them more difficult to keep suspended. Metal powder slurries, like their ceramic counterparts, are usually based on aqueous systems to which a suspending agent (to increase the viscosity) and dispersants (to prevent particle flocculation) are added (Kennedy and Lin, 2011). Other additions may be needed to lower the surface tension, change the pH or facilitate gelation.

The advantages of using metal slurries as the basis for producing foams is that constraints imposed by poor compressibility are removed, shaping can be performed by gelling in simple rubber moulds and gas bubbles can be introduced into the system by whisking or through the addition of gas-forming agents (although this means that the foam has to be stabilised). Although not well established, there are likely to be limitations to the materials that can be used based on interactions between the metal and the solvent (corrosion or reaction) and the ability to form a stable slurry.

#### **2.2.2 Slurry coating of polyurethane foams**

Porous metals can be produced by a replication method using an open cell polyurethane (PU) foam or sponge as a template (Quadbeck et al., 2007). In this process the polymer foam is first coated with a metal powder slurry, usually performed by immersion or by spraying. Excess slurry is removed by squeezing the foam, often by passing it through rollers. Without this, the cells may become partially closed due to the formation of liquid films bridging the cell struts. After coating and drying, the template is removed by thermal degradation and the resulting, fragile metal skeleton is further heated to sinter the metal powder particles, forming a rigid cellular metal structure.

Porous Metals and Metal Foams Made from Powders 41

Fig. 10. Micrograph of a steel foam produced by the slip reaction foaming process (Angel et

Foams have been produced from ceramic slurries by introducing air into the slurry, in much the same way as whisking cream (Sepulveda, 1997), and this method has recently been translated to metal systems (Lin, 2011). In the same way as for ceramics, in order to stabilise the air bubbles that are introduced during whisking, a surfactant must be added to the metal powder slurry. Through sufficient aeration, a foam can be formed which can be poured into a shaped mould made from almost any material. Despite the addition of surfactant, drainage of the liquid from the network of pores does occur, inevitably leading to collapse of the foam. To preserve the foam structure, slurry systems are designed to either gel by heating (cellulose systems) or cooling (agarose systems) or be polymerised by the addition of an initiator (acrylamide systems). The foamed body is then dried, further heated to burn

The pore structures for these foams are surprisingly uniform given the simplicity of the process, showing round pores connected by small windows. Figure 11 shows a cross section through a stainless steel foam in the gelled and dried condition, which demonstrates good

Fig. 11. A cross section (left) of a gelled and dried foam before sintering and (right) the foam microstructure showing an open cell structure and (inset) the porous nature of the cell struts

al., 2004).

(Lin, 2011).

**2.2.4 Foaming by mechanical whisking** 

out the polymer and finally sintered to densify the matrix.

Figure 9 shows a sintered porous stainless steel structure alongside the polymer foam used in the replication process. The strut structure is also shown and it is clear from the cross section that the struts are hollow (due to the space vacated by the polymer) and the walls of the struts are porous. Open cell structures with cell sizes between 10–80 pores per inch (about 2.5-0.2 mm) can be produced with precise pore structures with total porosities as high as 96%. This total porosity also includes the porosity in the sintered metal struts, but open cell porosities as high as 90% can be achieved.

This type of approach can be used to make metallic hollow spheres (which are themselves sintered to produce cellular metals as was described earlier). To make hollow spheres, expanded polystyrene spheres are coated with a metal powder slurry and then sintered, during which the polystyrene is removed and the metal forms a dense metal shell (Andersen et al., 2000).

Fig. 9. Images (left) showing a PU foam and a stainless steel foam made by slurry coating and (right) the microstructure of the struts showing, inset, their hollow nature.

#### **2.2.3 Slip reaction foam sintering**

Metal powder slurries can also be foamed by the insitu generation of a gas. In the slip reaction foam sintering (SRFS) process, the slurry (or slip) contains additives which stabilize the slip during processing (Angel et al., 2004) and to this a solution of ortho-phosphoric acid in either water or alcohol is then added. The hydrogen generated by the metal-acid reaction creates bubbles in the slip causing it to foam. As the solvent evaporates during the drying process, the pores formed by the hydrogen bubbles, which were originally closed, turn to interconnected porosity and an open cell green part is obtained. Figure 10 shows the typical structure for a steel foam which has irregular primary pore sizes as large as 3.5 mm and secondary pores between 0.05-0.3 mm. After sintering the total porosity is typically 60%. Foams have been produced from both steel and aluminium powders but for high porosities the green strength is low and cracks form in the foamed material.

In a variation on this process the slurry is an aqueous polymer solution that has the ability to form a gel (Shimizu and Matsuzaki, 2007). Gelation of the slurry is carried out by a freezing and thawing process and after gelation, it is heated until the foaming agent (hexane) decomposes, forming a gas, causing the gel to expand. The resulting foam is then dried and sintered.

Figure 9 shows a sintered porous stainless steel structure alongside the polymer foam used in the replication process. The strut structure is also shown and it is clear from the cross section that the struts are hollow (due to the space vacated by the polymer) and the walls of the struts are porous. Open cell structures with cell sizes between 10–80 pores per inch (about 2.5-0.2 mm) can be produced with precise pore structures with total porosities as high as 96%. This total porosity also includes the porosity in the sintered metal struts, but

This type of approach can be used to make metallic hollow spheres (which are themselves sintered to produce cellular metals as was described earlier). To make hollow spheres, expanded polystyrene spheres are coated with a metal powder slurry and then sintered, during which the polystyrene is removed and the metal forms a dense metal shell

Fig. 9. Images (left) showing a PU foam and a stainless steel foam made by slurry coating

Metal powder slurries can also be foamed by the insitu generation of a gas. In the slip reaction foam sintering (SRFS) process, the slurry (or slip) contains additives which stabilize the slip during processing (Angel et al., 2004) and to this a solution of ortho-phosphoric acid in either water or alcohol is then added. The hydrogen generated by the metal-acid reaction creates bubbles in the slip causing it to foam. As the solvent evaporates during the drying process, the pores formed by the hydrogen bubbles, which were originally closed, turn to interconnected porosity and an open cell green part is obtained. Figure 10 shows the typical structure for a steel foam which has irregular primary pore sizes as large as 3.5 mm and secondary pores between 0.05-0.3 mm. After sintering the total porosity is typically 60%. Foams have been produced from both steel and aluminium powders but for high porosities

In a variation on this process the slurry is an aqueous polymer solution that has the ability to form a gel (Shimizu and Matsuzaki, 2007). Gelation of the slurry is carried out by a freezing and thawing process and after gelation, it is heated until the foaming agent (hexane) decomposes, forming a gas, causing the gel to expand. The resulting foam is then

and (right) the microstructure of the struts showing, inset, their hollow nature.

the green strength is low and cracks form in the foamed material.

open cell porosities as high as 90% can be achieved.

(Andersen et al., 2000).

**2.2.3 Slip reaction foam sintering** 

dried and sintered.

#### **2.2.4 Foaming by mechanical whisking**

Foams have been produced from ceramic slurries by introducing air into the slurry, in much the same way as whisking cream (Sepulveda, 1997), and this method has recently been translated to metal systems (Lin, 2011). In the same way as for ceramics, in order to stabilise the air bubbles that are introduced during whisking, a surfactant must be added to the metal powder slurry. Through sufficient aeration, a foam can be formed which can be poured into a shaped mould made from almost any material. Despite the addition of surfactant, drainage of the liquid from the network of pores does occur, inevitably leading to collapse of the foam. To preserve the foam structure, slurry systems are designed to either gel by heating (cellulose systems) or cooling (agarose systems) or be polymerised by the addition of an initiator (acrylamide systems). The foamed body is then dried, further heated to burn out the polymer and finally sintered to densify the matrix.

The pore structures for these foams are surprisingly uniform given the simplicity of the process, showing round pores connected by small windows. Figure 11 shows a cross section through a stainless steel foam in the gelled and dried condition, which demonstrates good

Fig. 11. A cross section (left) of a gelled and dried foam before sintering and (right) the foam microstructure showing an open cell structure and (inset) the porous nature of the cell struts (Lin, 2011).

Porous Metals and Metal Foams Made from Powders 43

Fig. 13. Images of (left) a foam produced by the expansion of a melted PM precursor in a mould, (centre) an X-ray radiograph of the foam structure and (right) a cross section of the

**3 mm**

The foaming of compacted powder precursors has several key advantages. It is one of the few processes that can make closed cell foams, with the ability to produce near-net-shape and complex foam parts, including foam plates and sandwich structures. Lightweight Al foam parts made in this way are being used in a number of applications where their high strength at low mass and excellent energy absorbing ability are exploited. The main disadvantages are that large 3D parts are difficult to make and, in part due to the rapid nature of the foam expansion process, the uniformity and reproducibility of the pore structure can be unsatisfactory, leading to concerns over the reproducibility of the

Limits to the range of metals that can be foamed in this way arise due to several factors. There is a requirement for the blowing agent to only start to decompose when the metal or alloy is semi-solid. If decomposition occurs when the precursor is still solid, gas evolution can cause cracking of the precursor and escape of the gas without it contributing to foam expansion. It is, therefore, difficult to find suitable foaming agents for all metals. TiH2, which starts to decompose at around 450°C (Kennedy and Lopez, 2003), is used to foam low melting point metals, such as Al and Mg, despite decomposing below the melting point of most of their alloys. This requires the compacted precursor to contain very low (<2%) porosity to contain the gas (Kennedy, 2002). This is not readily achieved by conventional PM compaction methods and so to achieve the target densities, the use of high strength, prealloyed powders is avoided, favouring mixed elemental powder additions, and powder consolidation is either performed by hot die compaction, cold isostatic pressing followed by extrusion or continuous powder rolling or extrusion processes. The foaming of reactive metals and those with high melting points is not that practical, given the need for a conductive metal mould to produce shaped parts, but SrCO3 has been used to produce Fe-

Foams produced by melting compacted powders are stabilised (at least temporarily) by oxide films introduced into the liquid from the surface of the metal powders. The fraction of

**2.3.2 Advantages and limitations of the process** 

mechanical performance (Kennedy, 2004).

based foams.

foam showing the pore structure.

green strength, and also presents the sintered microstructures for the cells and the porous cell struts (inset). Porosities up to 90% have been achieved in sintered parts but work to date has shown that it is difficult to vary the pore size beyond the range of 0.5-1.5 mm. It is thought that this simple foaming process has the potential to produce foams that are suitable for both structural and functional applications.

#### **2.3 Foaming of compacted powder precursors**

#### **2.3.1 Foaming process**

Metal foams can be produced by encapsulating a foaming (or blowing) agent into a precursor made from compacted metal powder, followed by melting (Baumeister, 1990). The foaming agents are fine powdered compounds that, when heated, decompose to form a gas (typically they are metal hydrides or carbonates). When the compact is heated, usually in a mould, above the solidus temperature of the alloy, which should also be above the decomposition temperature of the foaming agent, the gas evolved causes expansion of the precursor. Expansion is rapid but collapse occurs, requiring fast cooling to "freeze-in" the foam structure. A schematic of the process used to make and foam powder precursors is shown in Figure 12.

Fig. 12. The sequence of steps used to manufacture metal foams made from compacted powder precursors (Ashby et al., 2000).

Figure 13 shows an expanded precursor that has been foamed in a metal mould, thereby defining its cylindrical shape. Also shown is a radiograph of the sample, revealing the porosity inside the part. The cross section of the foam, also shown in this figure, reveals that although the pores are reasonably spherical, a large variation in the pore sizes is observed. The closed porosity in foams made in this way is typically below 90%.

green strength, and also presents the sintered microstructures for the cells and the porous cell struts (inset). Porosities up to 90% have been achieved in sintered parts but work to date has shown that it is difficult to vary the pore size beyond the range of 0.5-1.5 mm. It is thought that this simple foaming process has the potential to produce foams that are

Metal foams can be produced by encapsulating a foaming (or blowing) agent into a precursor made from compacted metal powder, followed by melting (Baumeister, 1990). The foaming agents are fine powdered compounds that, when heated, decompose to form a gas (typically they are metal hydrides or carbonates). When the compact is heated, usually in a mould, above the solidus temperature of the alloy, which should also be above the decomposition temperature of the foaming agent, the gas evolved causes expansion of the precursor. Expansion is rapid but collapse occurs, requiring fast cooling to "freeze-in" the foam structure. A schematic of the process used to make and foam powder precursors is

Fig. 12. The sequence of steps used to manufacture metal foams made from compacted

The closed porosity in foams made in this way is typically below 90%.

Figure 13 shows an expanded precursor that has been foamed in a metal mould, thereby defining its cylindrical shape. Also shown is a radiograph of the sample, revealing the porosity inside the part. The cross section of the foam, also shown in this figure, reveals that although the pores are reasonably spherical, a large variation in the pore sizes is observed.

suitable for both structural and functional applications.

**2.3 Foaming of compacted powder precursors** 

**2.3.1 Foaming process** 

shown in Figure 12.

powder precursors (Ashby et al., 2000).

Fig. 13. Images of (left) a foam produced by the expansion of a melted PM precursor in a mould, (centre) an X-ray radiograph of the foam structure and (right) a cross section of the foam showing the pore structure.

#### **2.3.2 Advantages and limitations of the process**

The foaming of compacted powder precursors has several key advantages. It is one of the few processes that can make closed cell foams, with the ability to produce near-net-shape and complex foam parts, including foam plates and sandwich structures. Lightweight Al foam parts made in this way are being used in a number of applications where their high strength at low mass and excellent energy absorbing ability are exploited. The main disadvantages are that large 3D parts are difficult to make and, in part due to the rapid nature of the foam expansion process, the uniformity and reproducibility of the pore structure can be unsatisfactory, leading to concerns over the reproducibility of the mechanical performance (Kennedy, 2004).

Limits to the range of metals that can be foamed in this way arise due to several factors. There is a requirement for the blowing agent to only start to decompose when the metal or alloy is semi-solid. If decomposition occurs when the precursor is still solid, gas evolution can cause cracking of the precursor and escape of the gas without it contributing to foam expansion. It is, therefore, difficult to find suitable foaming agents for all metals. TiH2, which starts to decompose at around 450°C (Kennedy and Lopez, 2003), is used to foam low melting point metals, such as Al and Mg, despite decomposing below the melting point of most of their alloys. This requires the compacted precursor to contain very low (<2%) porosity to contain the gas (Kennedy, 2002). This is not readily achieved by conventional PM compaction methods and so to achieve the target densities, the use of high strength, prealloyed powders is avoided, favouring mixed elemental powder additions, and powder consolidation is either performed by hot die compaction, cold isostatic pressing followed by extrusion or continuous powder rolling or extrusion processes. The foaming of reactive metals and those with high melting points is not that practical, given the need for a conductive metal mould to produce shaped parts, but SrCO3 has been used to produce Febased foams.

Foams produced by melting compacted powders are stabilised (at least temporarily) by oxide films introduced into the liquid from the surface of the metal powders. The fraction of

Porous Metals and Metal Foams Made from Powders 45

Andersen O, Stephani G. 1999. Melt extracted fibres boost porous parts, *Metal Powder Report*,

Andersen O, Waag U, Schneider L, Stephani G, Kieback B. 2000. Novel metallic hollow

Andersen O, Studnitzky T, Bauer J. 2004. Direct typing: a new method for the production of

Asavavisithchai S., Kennedy A.R., 2006. Effect of powder oxide content on the expansion

Asavavisithchai S., Kennedy A.R., 2006. The Role of Oxidation During Compaction on the

Ashby MF, Evans A, Fleck NA, Gibson LJ, Hutchinson JW, Wadley HNG. 2000, *Metal Foams:* 

Banhart, J., 2001. Manufacture, characterisation and application of cellular metals and metal

Eisenmann M. 1998. Metal powder technologies and applications. In: *ASM Handbook, vol. 7.*

Jinnapat A, Kennedy A.R, 2010. The manufacture of spherical salt beads and their use as

Kanetake N., Kobashi M., 2006. Innovative processing of porous and cellular materials by

Kennedy A.R., 2002. The effect of compaction density on the foamability of Al-TiH2 powder

Kennedy A.R., Lopez V.H., 2003. The decomposition behavior of as-received and oxidized

Kennedy A.R., 2004, Aspects of the reproducibilty of mechanical properties in Al based

Kennedy, A R, Lin, X, 2011. Preparation and characterisation of metal powder slurries for

Lin X, 2011. Foaming of stainless steel powder slurries, *PhD thesis*, University of

Quadbeck P, Stephani G, Kuemmel K, Adler J, Standke G. 2007 Synthesis and properties of

open-celled metal foams. *Mater Sci Forum*, (534-536): 1005-1008.

use as precursors for metal foams made by gel casting *Powder Metallurgy,* 54,3, 376-

TiH2 foaming-agent powder, *Mat Sci and Eng*, A357, 1-2, 258-263.

dissolvable templates for the production of cellular solids via a powder metallurgy

Elzey DM, Wadley HNG. 2001. The Limits of Solid State Foaming*, Acta Mater*, 49, 849. Goodall, R., Mortensen, A., 2007. Microcellular aluminium: Child's play. *Advanced* 

cellular P/M parts. In: Danninger H, Ratzi R, editors. *Euro PM2004 Conference Proceedings*. Shrewsbury: European Powder Metallurgy Association , vol 4. p.189 Angel S, Bleck W, Scholz PF, Fend T. 2004. Influence of powder morphology and chemical

composition on metallic foams produced by Slip Reaction Foam Sintering (SRFS)-

and stability of PM-route Al foams, *Journal of Colloid and Interface Science*, 297, 715-

Expansion and Stability of Al Foams Made Via a PM Route, *Advanced Engineering* 

sphere structures, *Advanced Engineering Materials*, 2,192.

process. *Steel Res Intl*, 75, 483-488.

*A Design Guide*. Butterworth – Heinemann, UK.

Materials Park, USA: ASM International,1031.

route, *Journal of Alloys and Compounds*, 499, 43-47.

chemical reaction, Scripta Materialia 54, 521–525

Martin RL, Lederich RJ. 1992. Metal Powder Report, October, 30

compacts, *Powder Metallurgy*, 45, 1, 75-79

foams. *Progress in Materials Science*,. 46(6), 559-632.

**4. References** 

54, 30-34.

723.

*Materials*, 8, 568-572.

Baumeister J. 1990. German Patent 4,018,360,.

*Engineering Materials*, 9, 951-954.

foams, *J Mat Sci*, 39, 3085-3088.

381.

Nottingham, UK

these oxides in the expanding liquid is critical to achieving good foam structures and stable foams. Variations in oxide content for different powder sizes, from different suppliers or even for different batches of powders (due to variations in the atomising or storage conditions) can be the reason behind highly variable foaming responses (Asavavisithchai and Kennedy, 2006a). Varying the processing conditions during hot compaction can also affect the oxidation of the metal powder and alter the foaming behaviour (Asavavisithchai and Kennedy, 2006a, 2006b). Figure 14 shows the effect of oxygen content on the foam expansion for an Al powder, too low and foam collapse is severe, too high and the liquid is too viscous to foam. It should be noted that the way in which this oxygen (or oxide) level is achieved, through heat treatments at different temperatures or through atomisation, is not important.

Fig. 14. The effect of oxygen content on the foaming of pure Al powder compacts, showing that an optimum level is required (Asavavisithchai and Kennedy, 2006a).

#### **3. Summary**

An overview of some of the many and varied methods for making porous metals and metal foams from metal powders has been presented. With research and development into metal foams being vibrant and dynamic, pioneered by institutes like the Fraunhofer IFAM centres in Bremen, Chemnitz and Dresden in Germany, the state-of-the-art is continually evolving as the understanding behind powder processing, foaming and foam stabilisation improves.

For established foaming processes, research conducted within academia and industry has a strong emphasis on eliminating problems which would otherwise limit the wider use of these materials. This includes; improving the uniformity and reproducibility of the foam structures, aiming to achieve uniform pore sizes and densities throughout the component and similar foam structures from part to part; decreasing the processing and materials costs, through improved or new processing routes, reduced waste and cheaper starting materials and developing compelling case studies based on innovative design, simulation and testing to demonstrate to end users that despite higher prices for some foams or components containing foam elements, that these costs can be more that offset by the weight and energy savings offered by these novel materials and structures.

#### **4. References**

44 Powder Metallurgy

these oxides in the expanding liquid is critical to achieving good foam structures and stable foams. Variations in oxide content for different powder sizes, from different suppliers or even for different batches of powders (due to variations in the atomising or storage conditions) can be the reason behind highly variable foaming responses (Asavavisithchai and Kennedy, 2006a). Varying the processing conditions during hot compaction can also affect the oxidation of the metal powder and alter the foaming behaviour (Asavavisithchai and Kennedy, 2006a, 2006b). Figure 14 shows the effect of oxygen content on the foam expansion for an Al powder, too low and foam collapse is severe, too high and the liquid is too viscous to foam. It should be noted that the way in which this oxygen (or oxide) level is achieved, through heat treatments

0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

Oxygen content / wt.% Fig. 14. The effect of oxygen content on the foaming of pure Al powder compacts, showing

An overview of some of the many and varied methods for making porous metals and metal foams from metal powders has been presented. With research and development into metal foams being vibrant and dynamic, pioneered by institutes like the Fraunhofer IFAM centres in Bremen, Chemnitz and Dresden in Germany, the state-of-the-art is continually evolving as the understanding behind powder processing, foaming and foam stabilisation improves. For established foaming processes, research conducted within academia and industry has a strong emphasis on eliminating problems which would otherwise limit the wider use of these materials. This includes; improving the uniformity and reproducibility of the foam structures, aiming to achieve uniform pore sizes and densities throughout the component and similar foam structures from part to part; decreasing the processing and materials costs, through improved or new processing routes, reduced waste and cheaper starting materials and developing compelling case studies based on innovative design, simulation and testing to demonstrate to end users that despite higher prices for some foams or components containing foam elements, that these costs can be more that offset by the weight and energy

 heat treated at 500°C heat treated at 550°C

as received

that an optimum level is required (Asavavisithchai and Kennedy, 2006a).

at different temperatures or through atomisation, is not important.

0

savings offered by these novel materials and structures.

100

200

Maximum expansion / %

**3. Summary** 

300

400

500


**3** 

*Portugal* 

**Aluminium Alloy Foams:** 

Isabel Duarte and Mónica Oliveira *Centro de Tecnologia Mecânica e Automação,* 

**Production and Properties** 

*Departamento de Engenharia Mecânica, Universidade de Aveiro,* 

Ultra-light metal foams became an attractive research field both from the scientific and industrial applications view points. Closed-cell metal foams, in particular aluminium alloy (Al-alloy) ones can be used as lightweight, energy-absorption and damping structures in different industrial sectors, detaining an enormous potential when transportation is concerned. Despite the several manufacturing methods available, ultra-light metal foam applications seem to be restricted to a rather less demanding market in what concerns final product quality. The current available manufacturing processes enable effective control of density through process parameters manipulation. However, none of them allow for appropriate control of the cellular structures during its formation, leading to severe drawbacks in what concerns final product structural and mechanical properties. The latter, is undoubtedly the main reason for the lack of commercial acceptance of these ultra-light metal foams in product quality highly demanding sectors, such as the automotive or aeronautical sectors. The resolution of this problem is the main challenge of the scientific community in this field. To accomplish the latter two approaches may be followed: i) to develop new manufacturing processes or modify the existing ones to obtain foams with more uniform cellular structures. (ii) to understand and quantify the thermo-physicochemical mechanisms involved during the foam formation in order to control the process,

Metal foams manufacturing processes seem to abound (Banhart, 2001) and can be classified in two groups (Banhart, 2006): direct and indirect foaming methods. Direct foaming methods start from a molten metal containing uniformly dispersed ceramic particles to which gas bubbles are injected directly (Körner et al, 2005), or generated chemically by the decomposition of a blowing agent (e.g. titanium hydride, calcium), or by precipitation of gas dissolved in the melt by controlling temperature and pressure (Zeppelin, 2003). The indirect foaming methods require the preparation of foamable precursors that are subsequently foamed by heating. The foamable precursor consists of a dense compacted of powders

where the blowing agent particles are uniformly distributed into the metallic matrix.

Most commercially available metal foams are based on alloys containing (Degischer & Kirst, 2002): aluminium, nickel, magnesium, lead, copper, titanium, steel and even gold. Among

avoiding the occurrence of such imperfections and structural defects.

**1. Introduction**

Sepulveda P. 1997. Gelcasting foams for porous ceramics. *Am Ceram Soc Bull*, 76, 61-65.


## **Aluminium Alloy Foams: Production and Properties**

Isabel Duarte and Mónica Oliveira

*Centro de Tecnologia Mecânica e Automação, Departamento de Engenharia Mecânica, Universidade de Aveiro, Portugal* 

#### **1. Introduction**

46 Powder Metallurgy

Zhao, Y.Y., Sun D.X., 2001. A novel sintering-dissolution process for manufacturing Al

Zhou, J., 2006. *Advanced structural materials*. Porous Metallic Materials, ed. e. Winston O.

Sepulveda P. 1997. Gelcasting foams for porous ceramics. *Am Ceram Soc Bull*, 76, 61-65. Shimizu T, Matsuzaki K. 2007. Metal foam production process using hydro-gel and its

improvement. *Mater Sci Forum*, (539-543): 1845-1850.

Soboyejo., CRC Press , Taylor & Francis Group. USA, 22.

foams. *Scripta Materialia*,. 44, 1, 105-110.

Ultra-light metal foams became an attractive research field both from the scientific and industrial applications view points. Closed-cell metal foams, in particular aluminium alloy (Al-alloy) ones can be used as lightweight, energy-absorption and damping structures in different industrial sectors, detaining an enormous potential when transportation is concerned. Despite the several manufacturing methods available, ultra-light metal foam applications seem to be restricted to a rather less demanding market in what concerns final product quality. The current available manufacturing processes enable effective control of density through process parameters manipulation. However, none of them allow for appropriate control of the cellular structures during its formation, leading to severe drawbacks in what concerns final product structural and mechanical properties. The latter, is undoubtedly the main reason for the lack of commercial acceptance of these ultra-light metal foams in product quality highly demanding sectors, such as the automotive or aeronautical sectors. The resolution of this problem is the main challenge of the scientific community in this field. To accomplish the latter two approaches may be followed: i) to develop new manufacturing processes or modify the existing ones to obtain foams with more uniform cellular structures. (ii) to understand and quantify the thermo-physicochemical mechanisms involved during the foam formation in order to control the process, avoiding the occurrence of such imperfections and structural defects.

Metal foams manufacturing processes seem to abound (Banhart, 2001) and can be classified in two groups (Banhart, 2006): direct and indirect foaming methods. Direct foaming methods start from a molten metal containing uniformly dispersed ceramic particles to which gas bubbles are injected directly (Körner et al, 2005), or generated chemically by the decomposition of a blowing agent (e.g. titanium hydride, calcium), or by precipitation of gas dissolved in the melt by controlling temperature and pressure (Zeppelin, 2003). The indirect foaming methods require the preparation of foamable precursors that are subsequently foamed by heating. The foamable precursor consists of a dense compacted of powders where the blowing agent particles are uniformly distributed into the metallic matrix.

Most commercially available metal foams are based on alloys containing (Degischer & Kirst, 2002): aluminium, nickel, magnesium, lead, copper, titanium, steel and even gold. Among

Aluminium Alloy Foams: Production and Properties 49

Nowadays, foams manufactured by indirect foaming methods are also in the state of commercial exploitation, but in small-scale by German and Austrian Companies, like Schunk GmbH, Applied Light-weight Materials ALM and Austrian Company Alulight GmbH (Banhart, 2006). Powder Metallurgical (PM) method is one of the commercially exploited indirect methods to produce Al-alloy foams and it is also the research field of the authors of this chapter. This process consists on the heating of a precursor material which is obtained by hot compaction of a metal alloy (e.g. Al-alloy) with blowing agent powders (e.g. TiH2), resulting in the foam itself. The metal expands, developing a highly internal porous structure of closed-cells due to the simultaneous occurrence of the melting of the metal and thermal decomposition of the blowing agent with the release of a gas (e.g. H2). The liquid foam is then cooled in air, resulting in a solid foam with closed cells and with a very thin dense skin that improves the mechanical properties of these materials. This process can

The PM method has several advantages in comparison with the methods described earlier and will be further discussed ahead. The latter, has been addressed particularly in what concerns two research lines: (i) the study of the physics and foaming technology, with particular emphasis on dedicated process equipment development towards high quality foams production. (ii) foam quality assessment through proper part property characterisation, establishing its limits of application and seeking for new markets. This chapter presents a detailed overview of the current state-of-art in what concerns to methods, equipment and appropriate industrial procedures in order to obtain metal foams with good quality. The advantages, the disadvantages and the limitations of this PM method are also presented and discussed. An overview of the main challenges and perspectives in this field concerning industrial implementation is also presented, whenever it is possible the authors

Metal foam production by PM method can be divided into two production steps: (i) production of foamable precursor and (ii) production of the metal foam itself through the foaming of precursor material. A schematic diagram of the PM method is shown in Fig. 1. The first step is the preparation of a dense solid semi finished product called foamable precursor. The latter is attained by compacting a powder mixture containing the blowing agent and the metal, by using a conventional technique. The second step includes the production of the metal foam by heating this foamable precursor at temperatures above its

This PM method can be used to produce foams of different metals and its alloys (Degischer &Kriszt, 2002), such as aluminium and its alloys, tin, zinc, lead, steel and gold, which is one of the advantage of this PM method. Among all metal foams, the Al-alloys are the ones that have received more attention from both the research community and industry, due to its enormous potential mainly in what concerns specific weight and highly corrosion resistance. The most studied Al-alloys for foaming are pure aluminium, wrought alloys (e.g. 6xxx alloy series) or casting alloys (e.g. AlSi7Mg). The high quality foams of these different metals can be obtained by choosing the appropriate blowing agent. Moreover, the manufacturing parameters of the different stages require appropriate adjustment (Fig. 2).

produce foams with porosities between 75% and 90%.

present novel research work.

melting temperature.

**2. Production of metal foams** 

the metal foams, Al-alloys are commercially the most exploited ones due to their low density, high ductility, high thermal conductivity and competitive cost. Some manufacturing methods are already being commercialised. Direct foaming methods are currently being commercial exploited in a large-scale, the following companies are just a few examples. The Cymat Aluminium Corporation (in Canada) manufactures aluminium foams, designated "stabilised aluminium foam" which is obtained by gas injected directly into a molten metal (Degischer & Kirszt, 2002). Ceramic particles (e.g. silicon carbide, aluminium oxide and magnesium oxide) are used to enhance the viscosity of the melt and to adjust its foaming properties, since liquid metals cannot easily be foamed by the introduction of bubbling air. The foamy mass is relatively stable owing to the presence of ceramic particles in the melt. Foam panels with 1m in width and thickness range of 25-150 mm, can be produced continuously without length limitations at production rates of 900 kg/hour. The relative density of these foams is within the range 0.05-0.55 g/cm3. The average cell size is 2.5-30 mm. This process is the cheapest of all and allows for manufacturing large volume of foams. Moreover, through this process it is possible to obtain low density foams. The main disadvantage lies is the poor quality of the foams produced. The cell size is large and often irregular, and the foams tend to have a marked density gradient. Despite this process continuous improvement, the drawing of the foam and the size distribution of the pores are still difficult to control. Besides its low production cost, secondary operations are usually necessary. For example, the foamed material is cut into the required shape after foaming. The machining of these foams can be problematic due the high content of ceramic particles (10-30 vol. %) used in the process. Hütte Klein-Reichenbach Ges.m.b.H company (Austria) also produces and commercialises aluminium foams with excellent cell size uniformity, called MetComb. The process used is based on the gas injection method (Banhart, 2006) and allows for the production of complex shaped parts by casting the formed foam into the moulds.

An alternative way for foaming melts directly is to add a blowing agent to the molten metal. The blowing agent decomposes under the influence of heat and releases gas which then propels the foaming process. Shinko Wire Company has been manufacturing foamed aluminium under the registered trade name "Alporas" with production volumes reported as up to 1000 kg of foam per day, using a batch casting process (Miyoshi et al, 2000). This method starts with the addition of 1.5 wt.% calcium metal into the molten aluminium at 680°C, followed by several minutes stirring to adjust viscosity. An increase of viscosity is achieved by the formation of calcium oxides. After the viscosity has reached the desired value, titanium hydride (TiH2) is added (typically 1.6 wt.%), as a blowing agent by releasing hydrogen (H2) gas in the hot viscous liquid. The melt starts to expand slowly and gradually fills the foaming vessel. The foaming takes place at constant pressure. After cooling the vessel below the melting point of the alloy, the liquid foam turns into a solid Al foam. After that, the foam block is removed from the mould, it is sliced into flat plates of various thicknesses according to its end use. This process is capable of producing large blocks of good quality. Blocks with 450 mm in width, 2050 mm in length and 650 mm in height can be produced. These foams have uniform pore structure and do not require the addition of ceramic particles, which makes it brittle. However, the method is more expensive than foaming melts by gas injection method requiring more complex processing equipment. The density range of these foams is 0.18-0.24 g/cm3, and the mean cell size is about 4.5 mm.

the metal foams, Al-alloys are commercially the most exploited ones due to their low density, high ductility, high thermal conductivity and competitive cost. Some manufacturing methods are already being commercialised. Direct foaming methods are currently being commercial exploited in a large-scale, the following companies are just a few examples. The Cymat Aluminium Corporation (in Canada) manufactures aluminium foams, designated "stabilised aluminium foam" which is obtained by gas injected directly into a molten metal (Degischer & Kirszt, 2002). Ceramic particles (e.g. silicon carbide, aluminium oxide and magnesium oxide) are used to enhance the viscosity of the melt and to adjust its foaming properties, since liquid metals cannot easily be foamed by the introduction of bubbling air. The foamy mass is relatively stable owing to the presence of ceramic particles in the melt. Foam panels with 1m in width and thickness range of 25-150 mm, can be produced continuously without length limitations at production rates of 900 kg/hour. The relative density of these foams is within the range 0.05-0.55 g/cm3. The average cell size is 2.5-30 mm. This process is the cheapest of all and allows for manufacturing large volume of foams. Moreover, through this process it is possible to obtain low density foams. The main disadvantage lies is the poor quality of the foams produced. The cell size is large and often irregular, and the foams tend to have a marked density gradient. Despite this process continuous improvement, the drawing of the foam and the size distribution of the pores are still difficult to control. Besides its low production cost, secondary operations are usually necessary. For example, the foamed material is cut into the required shape after foaming. The machining of these foams can be problematic due the high content of ceramic particles (10-30 vol. %) used in the process. Hütte Klein-Reichenbach Ges.m.b.H company (Austria) also produces and commercialises aluminium foams with excellent cell size uniformity, called MetComb. The process used is based on the gas injection method (Banhart, 2006) and allows for the production of complex shaped parts by casting the formed foam into the

An alternative way for foaming melts directly is to add a blowing agent to the molten metal. The blowing agent decomposes under the influence of heat and releases gas which then propels the foaming process. Shinko Wire Company has been manufacturing foamed aluminium under the registered trade name "Alporas" with production volumes reported as up to 1000 kg of foam per day, using a batch casting process (Miyoshi et al, 2000). This method starts with the addition of 1.5 wt.% calcium metal into the molten aluminium at 680°C, followed by several minutes stirring to adjust viscosity. An increase of viscosity is achieved by the formation of calcium oxides. After the viscosity has reached the desired value, titanium hydride (TiH2) is added (typically 1.6 wt.%), as a blowing agent by releasing hydrogen (H2) gas in the hot viscous liquid. The melt starts to expand slowly and gradually fills the foaming vessel. The foaming takes place at constant pressure. After cooling the vessel below the melting point of the alloy, the liquid foam turns into a solid Al foam. After that, the foam block is removed from the mould, it is sliced into flat plates of various thicknesses according to its end use. This process is capable of producing large blocks of good quality. Blocks with 450 mm in width, 2050 mm in length and 650 mm in height can be produced. These foams have uniform pore structure and do not require the addition of ceramic particles, which makes it brittle. However, the method is more expensive than foaming melts by gas injection method requiring more complex processing equipment. The density range of these foams is 0.18-0.24 g/cm3, and the mean cell size is about 4.5 mm.

moulds.

Nowadays, foams manufactured by indirect foaming methods are also in the state of commercial exploitation, but in small-scale by German and Austrian Companies, like Schunk GmbH, Applied Light-weight Materials ALM and Austrian Company Alulight GmbH (Banhart, 2006). Powder Metallurgical (PM) method is one of the commercially exploited indirect methods to produce Al-alloy foams and it is also the research field of the authors of this chapter. This process consists on the heating of a precursor material which is obtained by hot compaction of a metal alloy (e.g. Al-alloy) with blowing agent powders (e.g. TiH2), resulting in the foam itself. The metal expands, developing a highly internal porous structure of closed-cells due to the simultaneous occurrence of the melting of the metal and thermal decomposition of the blowing agent with the release of a gas (e.g. H2). The liquid foam is then cooled in air, resulting in a solid foam with closed cells and with a very thin dense skin that improves the mechanical properties of these materials. This process can produce foams with porosities between 75% and 90%.

The PM method has several advantages in comparison with the methods described earlier and will be further discussed ahead. The latter, has been addressed particularly in what concerns two research lines: (i) the study of the physics and foaming technology, with particular emphasis on dedicated process equipment development towards high quality foams production. (ii) foam quality assessment through proper part property characterisation, establishing its limits of application and seeking for new markets. This chapter presents a detailed overview of the current state-of-art in what concerns to methods, equipment and appropriate industrial procedures in order to obtain metal foams with good quality. The advantages, the disadvantages and the limitations of this PM method are also presented and discussed. An overview of the main challenges and perspectives in this field concerning industrial implementation is also presented, whenever it is possible the authors present novel research work.

#### **2. Production of metal foams**

Metal foam production by PM method can be divided into two production steps: (i) production of foamable precursor and (ii) production of the metal foam itself through the foaming of precursor material. A schematic diagram of the PM method is shown in Fig. 1. The first step is the preparation of a dense solid semi finished product called foamable precursor. The latter is attained by compacting a powder mixture containing the blowing agent and the metal, by using a conventional technique. The second step includes the production of the metal foam by heating this foamable precursor at temperatures above its melting temperature.

This PM method can be used to produce foams of different metals and its alloys (Degischer &Kriszt, 2002), such as aluminium and its alloys, tin, zinc, lead, steel and gold, which is one of the advantage of this PM method. Among all metal foams, the Al-alloys are the ones that have received more attention from both the research community and industry, due to its enormous potential mainly in what concerns specific weight and highly corrosion resistance. The most studied Al-alloys for foaming are pure aluminium, wrought alloys (e.g. 6xxx alloy series) or casting alloys (e.g. AlSi7Mg). The high quality foams of these different metals can be obtained by choosing the appropriate blowing agent. Moreover, the manufacturing parameters of the different stages require appropriate adjustment (Fig. 2).

Aluminium Alloy Foams: Production and Properties 51

already be manufactured (Degischer & Kirst, 2002). Furthermore, it must be pointed out that during PM method it is still rather difficult to fully control the foaming process, which

The first step of the PM method to obtain metal foams is the production of foamable precursor materials. The fundamental aspects of this production step are discussed ahead.

The blowing agent is a chemical compound which releases gas when heated, being the responsible for the formation of bubbles. There are two main requirements to obtain high quality foams. The first one is to ensure an uniform distribution of the blowing particles into the metal matrix within the precursor material. The second is to ensure the coordination of the thermal decomposition characteristics of the blowing agent and the alloy melting behaviour to avoid cracks' formation before melting. The selection of these powders is therefore, detrimental for the foaming success. The characteristics of the metal powders, like the purity, the particle size, the alloying chemical elements (content and type) and the impurities (content and type), as well as the alloy melting behaviour and the thermal decomposition characteristics of the blowing agent must be studied and known. The literature highlights how powders from different manufactures could lead to notable

differences in foaming behaviour (Baumgärtner et al, 2000; Degischer & Kirst, 2002).

The blowing agents usually used for producing Al-alloy foams using the PM method are metal hydrides, such as titanium hydride (TiH2), zirconium hydride (ZrH2) and magnesium hydride (MgH2) (Duarte & Banhart, 2000). The employment of other blowing agent powders, such as the carbonates, has been investigated as a cost-effective alternative to metal hydrides (Haesche et al, 2010; Cambronero et al, 2009). It seems, though, that titanium hydride still is the best choice for the foaming agent when producing Al-alloy foams, as reported elsewhere (Duarte & Banhart, 2000). Fig. 3 presents a typical mixture of Al-alloy and TiH2 powders. The amount of the metal hydrides usually used is less than 1% in weight in the initial powder mixture, based upon the Al or Al-alloy that is to be foamed. For example, high quality of AlSi7 foams can be obtained by using 0.6wt.% in the initial powder

The effect of the composition of the alloy, the impurities, the particle size and the alloying chemical elements on the foaming behaviour has been studied (Duarte & Banhart, 2000;

(a) (b)

results in lack of uniformity of the pore structure.

**2.1 Preparation of the foamable precursors** 

**2.1.1 Selection of the powders** 

mixture (Duarte & Banhart, 2000).

Fig. 3. Al-alloy and TiH2 powders used in PM method.

Fig. 1. Powder Metallurgical metod for making metal foams.

Fig. 2. Manufacturing parameters of PM method (Duarte, 2005).

The main advantage of this PM method is to enable the production of components of metal foams with different architectures (e.g. sandwich systems, filled profiles and 3D complex shaped structures) in comparison with the others (Duarte et al, 2008, 2010). The materials can be joined during the foaming step without using chemical adhesives (Duarte et al, 2006). Other advantage lies in the fact that the addition of ceramic particles are not required, avoiding the brittle mechanical behaviour inferred by these particles. Moreover, the foam parts are covered by an external dense metal skin that improves its mechanical behaviour, providing a good surface finish. The disadvantage of this PM process is the high production cost mainly associated to the powder prices. Another disadvantage is the difficulty to manufacture large volume foam parts. Nevertheless, sandwich panels of 2mx1mx1cm can already be manufactured (Degischer & Kirst, 2002). Furthermore, it must be pointed out that during PM method it is still rather difficult to fully control the foaming process, which results in lack of uniformity of the pore structure.

#### **2.1 Preparation of the foamable precursors**

The first step of the PM method to obtain metal foams is the production of foamable precursor materials. The fundamental aspects of this production step are discussed ahead.

#### **2.1.1 Selection of the powders**

50 Powder Metallurgy

First production step: Production of foamable precursor

Second production step: Production of metal foam

Cold isostatic pressing

extruding

Bonding with metal sheets

rolling

Foamable rods

Foamable sandwiches

3-D foam parts (any geometries including complex geometry)

Hollow structures filled by foam

3-D foam parts with metallic inserts

Sandwich panels

Foamable sheets

Cold and hot pressing Foamable tablets

Foaming (into batch or continuous foaming furnaces)

Fig. 1. Powder Metallurgical metod for making metal foams.

Closed-mould

Open-mould

Blowing agent

Mixing

Mixture

Al-alloy

Foamable precursors (tablets, rods and sheets)

Fig. 2. Manufacturing parameters of PM method (Duarte, 2005).

The main advantage of this PM method is to enable the production of components of metal foams with different architectures (e.g. sandwich systems, filled profiles and 3D complex shaped structures) in comparison with the others (Duarte et al, 2008, 2010). The materials can be joined during the foaming step without using chemical adhesives (Duarte et al, 2006). Other advantage lies in the fact that the addition of ceramic particles are not required, avoiding the brittle mechanical behaviour inferred by these particles. Moreover, the foam parts are covered by an external dense metal skin that improves its mechanical behaviour, providing a good surface finish. The disadvantage of this PM process is the high production cost mainly associated to the powder prices. Another disadvantage is the difficulty to manufacture large volume foam parts. Nevertheless, sandwich panels of 2mx1mx1cm can The blowing agent is a chemical compound which releases gas when heated, being the responsible for the formation of bubbles. There are two main requirements to obtain high quality foams. The first one is to ensure an uniform distribution of the blowing particles into the metal matrix within the precursor material. The second is to ensure the coordination of the thermal decomposition characteristics of the blowing agent and the alloy melting behaviour to avoid cracks' formation before melting. The selection of these powders is therefore, detrimental for the foaming success. The characteristics of the metal powders, like the purity, the particle size, the alloying chemical elements (content and type) and the impurities (content and type), as well as the alloy melting behaviour and the thermal decomposition characteristics of the blowing agent must be studied and known. The literature highlights how powders from different manufactures could lead to notable differences in foaming behaviour (Baumgärtner et al, 2000; Degischer & Kirst, 2002).

The blowing agents usually used for producing Al-alloy foams using the PM method are metal hydrides, such as titanium hydride (TiH2), zirconium hydride (ZrH2) and magnesium hydride (MgH2) (Duarte & Banhart, 2000). The employment of other blowing agent powders, such as the carbonates, has been investigated as a cost-effective alternative to metal hydrides (Haesche et al, 2010; Cambronero et al, 2009). It seems, though, that titanium hydride still is the best choice for the foaming agent when producing Al-alloy foams, as reported elsewhere (Duarte & Banhart, 2000). Fig. 3 presents a typical mixture of Al-alloy and TiH2 powders. The amount of the metal hydrides usually used is less than 1% in weight in the initial powder mixture, based upon the Al or Al-alloy that is to be foamed. For example, high quality of AlSi7 foams can be obtained by using 0.6wt.% in the initial powder mixture (Duarte & Banhart, 2000).

The effect of the composition of the alloy, the impurities, the particle size and the alloying chemical elements on the foaming behaviour has been studied (Duarte & Banhart, 2000;

Fig. 3. Al-alloy and TiH2 powders used in PM method.

Aluminium Alloy Foams: Production and Properties 53

Another strategy is to change the alloy composition to obtain lower melting temperatures through the addition of alloying elements, such as the magnesium, zinc and copper that decreases the melting temperature (Helwig et al, 2011). Although this strategy appears to be promising, research in this field has not been very systematic. The results revealed that these treatments form a sufficiently thick oxide layer which leads to a minimum of hydrogen loose. Helwig *et al* reported that the use of even higher magnesium amounts was found out to lead to promising results in the run-up to this study. Researchers have been testing the addition of the ceramic particles (e.g. alumina) in the initial powder mixture to improve the foaming stability through the increase of the melt bulk viscosity (Kennedy, 2004). However, the presence of these particles can originate a brittle mechanical behaviour of the foams.

The mixing procedure should yield a homogeneous distribution of the alloying elements and the blowing agent particles to ensure the high-quality of the Al-alloy foams with an uniform pore size distribution. Blending the Al-alloy and the blowing agent powders is a crucial step within the entire foaming process. This operation should be made to avoid the agglomeration of the blowing agent particles and the alloying elements. This causes structural defects and imperfections on the final foam. The mixers usually used are tumbling mixers with or without alumina balls (Fig. 5). These balls do not add any other element to

An important practical aspect in the mixing operation is to obtain a clean and homogeneous powder mixture. The impurities and solid powders by dirt, water or other particles entrapped in the mixture may have a detrimental effect in foaming. These impurities can act like nuclei uncontrolled voids during the thermal decomposition of the blowing agent which will form larger pores at the latest foaming stages (Matijasevic & Banhart, 2006).

(a) (b)

The compaction should ensure that the blowing agent particles are embedded in the Alalloy matrix. The basic practical rule is to obtain a dense semi compact called foamable precursor material with no residual open porosity in which theoretical density is close to 100% of the theoretical density of the aluminium matrix (Duarte&Banhart, 2000). The production of precursors by compacting powder mixtures can be performed in a variety of ways, e.g., by uni-axial (Kennedy, 2004), double-axial or isostatic pressing (Körner et al, 2000), extrusion (Baumgartner et al, 2000), powder rolling (Kitazono, 2004), etc., and all the

Fig. 5. (a) Turbula mixer used to mix the powders. (b) Powder mixtures.

the mixture, because the aluminium oxide is already present in the mixture.

**2.1.2 Mixing of the powders** 

**2.1.3 Compaction of the powders** 

Lehmus & Busse, 2004; Gokhale et al, 2007; Ibrahim et al, 2008; Helwig et al, 2011). Some of these effects on the foam expansion behaviour or cell structure have not yet been well established. A recent research of the effect of TiH2 particle size on the foaming behaviour and on the morphology of Al-alloy foam produced by PM process is reported (Ibrahim et al, 2008). These studies revealed that the use of the coarser particle sizes of TiH2 leads to a higher foam expansion and coarser macrostructure while the finer grade of TiH2 leads to a quite lower maximum expansion and a finer macrostructure. The use of different particle sizes is an approach to adapt the onset of gas evolution temperature of the gas blowing agent and improvement of the macrostructure of foamed aluminium.

The difference between thermal decomposition of the blowing agent and melting temperature of the metal may cause the formation of irregular, crack-like pores in early expansion stages, which then can lead to irregularities in the final foam (Duarte & Banhart, 2000). The research in this field has been demonstrating that the high-quality Al-alloy foam is obtained when this difference is minimised. The basic rules to choose the best blowing agent are related to the closing between the temperature of the beginning of the thermal decomposition of the blowing agent and the *liquidus* temperature of the metal. This problem has been approached in two different ways: (i) pre-treatments of the blowing agent powder to delay the hydrogen release, i.e. to shift it to higher temperatures, (ii) to change the alloy composition to obtain lower melting point through the addition of alloying elements.

Thermal pre-treatments that lead to partial decomposition and/or pre-oxidation of the powder surface (Matijašević et al, 2006) or surface coatings (Proa-Flores & Drew, 2008) have been applied. For example, TiH2 powder particles are heated (480ºC) in air during a given time (180 min) and an oxide layer is formed on the surface of the particles. This layer delays gas release from the particles, so that hydrogen is ideally released during foaming only after the alloy melting temperature has been reached. The powder colour depends on the thickness of the formed oxide layer (Fig. 4). Another example, the TiH2 powders can be treated with acetic acid solution during 10 h, followed by a wash with distilled water until the solution presented a low acidity. The particles can then be coated with a silicon dioxide layer (Fang et al, 2005).

Fig. 4. Untreated (a) and treated (b) TiH2 powders and their TG/DTG curves (c).

Another strategy is to change the alloy composition to obtain lower melting temperatures through the addition of alloying elements, such as the magnesium, zinc and copper that decreases the melting temperature (Helwig et al, 2011). Although this strategy appears to be promising, research in this field has not been very systematic. The results revealed that these treatments form a sufficiently thick oxide layer which leads to a minimum of hydrogen loose. Helwig *et al* reported that the use of even higher magnesium amounts was found out to lead to promising results in the run-up to this study. Researchers have been testing the addition of the ceramic particles (e.g. alumina) in the initial powder mixture to improve the foaming stability through the increase of the melt bulk viscosity (Kennedy, 2004). However, the presence of these particles can originate a brittle mechanical behaviour of the foams.

#### **2.1.2 Mixing of the powders**

52 Powder Metallurgy

Lehmus & Busse, 2004; Gokhale et al, 2007; Ibrahim et al, 2008; Helwig et al, 2011). Some of these effects on the foam expansion behaviour or cell structure have not yet been well established. A recent research of the effect of TiH2 particle size on the foaming behaviour and on the morphology of Al-alloy foam produced by PM process is reported (Ibrahim et al, 2008). These studies revealed that the use of the coarser particle sizes of TiH2 leads to a higher foam expansion and coarser macrostructure while the finer grade of TiH2 leads to a quite lower maximum expansion and a finer macrostructure. The use of different particle sizes is an approach to adapt the onset of gas evolution temperature of the gas blowing

The difference between thermal decomposition of the blowing agent and melting temperature of the metal may cause the formation of irregular, crack-like pores in early expansion stages, which then can lead to irregularities in the final foam (Duarte & Banhart, 2000). The research in this field has been demonstrating that the high-quality Al-alloy foam is obtained when this difference is minimised. The basic rules to choose the best blowing agent are related to the closing between the temperature of the beginning of the thermal decomposition of the blowing agent and the *liquidus* temperature of the metal. This problem has been approached in two different ways: (i) pre-treatments of the blowing agent powder to delay the hydrogen release, i.e. to shift it to higher temperatures, (ii) to change the alloy

composition to obtain lower melting point through the addition of alloying elements.

Thermal pre-treatments that lead to partial decomposition and/or pre-oxidation of the powder surface (Matijašević et al, 2006) or surface coatings (Proa-Flores & Drew, 2008) have been applied. For example, TiH2 powder particles are heated (480ºC) in air during a given time (180 min) and an oxide layer is formed on the surface of the particles. This layer delays gas release from the particles, so that hydrogen is ideally released during foaming only after the alloy melting temperature has been reached. The powder colour depends on the thickness of the formed oxide layer (Fig. 4). Another example, the TiH2 powders can be treated with acetic acid solution during 10 h, followed by a wash with distilled water until the solution presented a low acidity. The particles can then be coated with a silicon dioxide

TG (%)

(a) (b) (c)

Fig. 4. Untreated (a) and treated (b) TiH2 powders and their TG/DTG curves (c).

TG: TiH2

TG: TiH2

DTG: TiH2

DTG: TiH2





0 200 400 600 800 1000

Temperature (ºC)

air atmosphere 10k/min

0,0 0,2 0,4 0,6 0,8 1,0 1,2 1,4 1,6 1,8

DTG (%/min)

agent and improvement of the macrostructure of foamed aluminium.

layer (Fang et al, 2005).

The mixing procedure should yield a homogeneous distribution of the alloying elements and the blowing agent particles to ensure the high-quality of the Al-alloy foams with an uniform pore size distribution. Blending the Al-alloy and the blowing agent powders is a crucial step within the entire foaming process. This operation should be made to avoid the agglomeration of the blowing agent particles and the alloying elements. This causes structural defects and imperfections on the final foam. The mixers usually used are tumbling mixers with or without alumina balls (Fig. 5). These balls do not add any other element to the mixture, because the aluminium oxide is already present in the mixture.

An important practical aspect in the mixing operation is to obtain a clean and homogeneous powder mixture. The impurities and solid powders by dirt, water or other particles entrapped in the mixture may have a detrimental effect in foaming. These impurities can act like nuclei uncontrolled voids during the thermal decomposition of the blowing agent which will form larger pores at the latest foaming stages (Matijasevic & Banhart, 2006).

Fig. 5. (a) Turbula mixer used to mix the powders. (b) Powder mixtures.

#### **2.1.3 Compaction of the powders**

The compaction should ensure that the blowing agent particles are embedded in the Alalloy matrix. The basic practical rule is to obtain a dense semi compact called foamable precursor material with no residual open porosity in which theoretical density is close to 100% of the theoretical density of the aluminium matrix (Duarte&Banhart, 2000). The production of precursors by compacting powder mixtures can be performed in a variety of ways, e.g., by uni-axial (Kennedy, 2004), double-axial or isostatic pressing (Körner et al, 2000), extrusion (Baumgartner et al, 2000), powder rolling (Kitazono, 2004), etc., and all the

Aluminium Alloy Foams: Production and Properties 55

(a) (b) (c) Fig. 8. Precursor materials at different densification levels: (a,b) low-quality (c) high-quality.

Several parameters are evaluated to ensure the quality of the foamable precursors. The main parameters are the density and the distribution of the blowing agent particles into the metal

matrix. The latter is usually evaluated by scanning electron microscope (Fig. 9).

Fig. 9. Foamable precursor sample of aluminium alloys containing 0.6 wt.% of TiH2

range of normal statistical fluctuations.

composed with Al-matrix (dark gray) with Si (light gray) and TiH2 (white colour) particles. The required density of the foamable precursor is adjusted by manipulating the compaction parameters (time and temperature). Duarte and Banhart reported that the hot-pressing temperature is a very important parameter (Duarte & Banhart, 2000). The reduction of compaction temperatures lead to insufficient compaction with some residual porosity (see density values given in Fig.10 a). The hydrogen gas can escape from the melting alloy without creating pores in this case. The same phenomenon is observed when the powders are extruded instead of hot pressed. On the other hand when higher compaction temperatures are employed lower maximum expansions can be observed, mainly because some of the hydrogen is lost during compaction. Even higher compaction temperatures lead to a rapid loss of foamability. The optimum compaction temperature therefore lies around 450°C for the case investigated, well above the initial decomposition temperature of TiH2 (380°C). The compaction time is not considered a critical parameter for the compaction temperatures chosen. The variations observed are not systematic and seem to be within the

**200m** 

above mentioned techniques can be hot or cold. A conjugation of different conventional compaction techniques can be used. Furthermore, the compacting process can be performed in an inert atmosphere, in air or in vacuum (Jiménez et al, 2009). The most economical way is the double-axial pressing in air, but the most efficient one is the high-temperature extrusion. There are several procedures to produce foamable precursors. A simple process is to compact the powder mixture using a hot uniaxial pressing at temperatures close to the thermal decomposition of the blowing agent, using a pressing device and a die with a heating system (Fig. 6). This process enables to yield more than 99% relative density of the precursor (Fig. 7a).

Fig. 6. Die with the heating system used to prepare the precursor material.

The initial powder mixture is first compacted to cylindrical billets of 70-80% of the theoretical density, using the cold isostatic pressing. Then, these billets are pre-heated to temperatures close to the thermal decomposition of the blowing agent and extruded into rectangular bars of various dimensions (Baumgärtner et al, 2000), as shown in Fig. 7b.

Kennedy reported that a minimum compaction density is required to achieve appreciable expansion and is about 94% of the theoretical density (Kennedy, 2002). The highest value of the expansion is obtained when the density is close to the theoretical density of the Al-alloy matrix. The authors research results reveal that compaction of the mixture to achieve high quality precursor materials must be performed at temperatures close to the initial thermal decomposition temperature (e.g. 400ºC), in order to achieve the density value which leads to a good foaming behaviour Moreover, the cold compaction should be done before the hotcompaction in order to obtain a high level of densification (Fig.8).

Fig. 7. Precursor materials of Al-alloy containing 0.6%wt. of TiH2 which is manufactured using the laboratory system in Fig.6 (a) and supplied by IFAM (Baumgärtner et al, 2000).

above mentioned techniques can be hot or cold. A conjugation of different conventional compaction techniques can be used. Furthermore, the compacting process can be performed in an inert atmosphere, in air or in vacuum (Jiménez et al, 2009). The most economical way is the double-axial pressing in air, but the most efficient one is the high-temperature extrusion. There are several procedures to produce foamable precursors. A simple process is to compact the powder mixture using a hot uniaxial pressing at temperatures close to the thermal decomposition of the blowing agent, using a pressing device and a die with a heating system (Fig. 6). This process enables to yield more than 99% relative density of the

Fig. 6. Die with the heating system used to prepare the precursor material.

compaction in order to obtain a high level of densification (Fig.8).

The initial powder mixture is first compacted to cylindrical billets of 70-80% of the theoretical density, using the cold isostatic pressing. Then, these billets are pre-heated to temperatures close to the thermal decomposition of the blowing agent and extruded into rectangular bars of various dimensions (Baumgärtner et al, 2000), as shown in Fig. 7b.

Kennedy reported that a minimum compaction density is required to achieve appreciable expansion and is about 94% of the theoretical density (Kennedy, 2002). The highest value of the expansion is obtained when the density is close to the theoretical density of the Al-alloy matrix. The authors research results reveal that compaction of the mixture to achieve high quality precursor materials must be performed at temperatures close to the initial thermal decomposition temperature (e.g. 400ºC), in order to achieve the density value which leads to a good foaming behaviour Moreover, the cold compaction should be done before the hot-

(a) (b)

Fig. 7. Precursor materials of Al-alloy containing 0.6%wt. of TiH2 which is manufactured using the laboratory system in Fig.6 (a) and supplied by IFAM (Baumgärtner et al, 2000).

precursor (Fig. 7a).

Fig. 8. Precursor materials at different densification levels: (a,b) low-quality (c) high-quality.

Several parameters are evaluated to ensure the quality of the foamable precursors. The main parameters are the density and the distribution of the blowing agent particles into the metal matrix. The latter is usually evaluated by scanning electron microscope (Fig. 9).

The required density of the foamable precursor is adjusted by manipulating the compaction parameters (time and temperature). Duarte and Banhart reported that the hot-pressing temperature is a very important parameter (Duarte & Banhart, 2000). The reduction of compaction temperatures lead to insufficient compaction with some residual porosity (see density values given in Fig.10 a). The hydrogen gas can escape from the melting alloy without creating pores in this case. The same phenomenon is observed when the powders are extruded instead of hot pressed. On the other hand when higher compaction temperatures are employed lower maximum expansions can be observed, mainly because some of the hydrogen is lost during compaction. Even higher compaction temperatures lead to a rapid loss of foamability. The optimum compaction temperature therefore lies around 450°C for the case investigated, well above the initial decomposition temperature of TiH2 (380°C). The compaction time is not considered a critical parameter for the compaction temperatures chosen. The variations observed are not systematic and seem to be within the range of normal statistical fluctuations. **200m** 

Aluminium Alloy Foams: Production and Properties 57

developed (Fig. 13 b). The disadvantage of the process is to produce components with a

(a) (b) (c) Fig. 12. Practical aspects of the process. (a) closed mould. (b) mould with the precursor into

(a) (b)

The atmosphere, heating rate and temperature of the thermal foaming cycle are some of the manufacturing parameters which influence the quality and the properties of the resulting foam. Moreover, it should also be referred that the characteristics of the mould (material, design and dimensions), the type of the furnace (batch or continuous) should also be considered (Duarte, 2005). The effects of these variables on the foaming behaviour may therefore be investigated. To assess the latter, foaming tests are performed by heating the

Fig. 13. Foaming furnaces. Batch chamber furnace (a). Continuous foaming furnace

large volume parts as it is the case of Shinko Wire method (Miyoshi et al, 2000).

the pre-heated furnace. (c) Al-alloy foam block.

developed by this research team (b).

Fig. 14. 3D-parts of Al-alloy foams.

Fig. 10. Effect of the hot pressing temperature on the precursor density (a) and on the foaming behaviour (b).

#### **2.2 Production of metal foams**

The second step is the production of metal foams by heating the foamable precursor at temperatures above the melting point of the alloy. The metal expands developing a highly porous closed-cell internal structure due to the simultaneous melting of the aluminium and thermal decomposition of the blowing agent (TiH2) in gas (H2). The solid foams are obtained by controlled cooling of the formed liquid foam, at temperatures below the *solidus* temperature. The foamed parts have a highly internal structure with closed cells (Fig. 11 b) and are covered by a dense metal skin that improves their mechanical properties and provide good surface finish (Fig. 11).

The foaming process usually takes place into the stainless steel closed moulds (Fig. 12). The cavity of the mould should have the same design and dimensions of the final foam. The foamable precursor into the mould is heated at temperatures above its melting point (Fig. 12 b). The material expands and fills the entire mould cavity. After that, the mould is cooled, followed by the extraction of the foam (Fig. 12c). The furnaces used to manufacture the metal foams usually are of the batch chamber furnace type (Fig. 13a). This research team has developed a foaming continuous furnace to produce these materials (Fig. 13b).

Fig. 11. Al-alloy foam covered by a dense skin (a), with a closed-cell internal structure (b).

This method enables the cost effective production of aluminium alloy parts without limitations concerning shape ( e.g. panels, profiles or complex 3D shaped parts) (Duarte et al, 2006, 2008, 2010). Here, the group research results are presented to illustrate some practical examples (Fig. 14) which can be obtained by using the continuous foaming furnace

Sample temperature [°C]

(a) (b)

The second step is the production of metal foams by heating the foamable precursor at temperatures above the melting point of the alloy. The metal expands developing a highly porous closed-cell internal structure due to the simultaneous melting of the aluminium and thermal decomposition of the blowing agent (TiH2) in gas (H2). The solid foams are obtained by controlled cooling of the formed liquid foam, at temperatures below the *solidus* temperature. The foamed parts have a highly internal structure with closed cells (Fig. 11 b) and are covered by a dense metal skin that improves their mechanical properties and

The foaming process usually takes place into the stainless steel closed moulds (Fig. 12). The cavity of the mould should have the same design and dimensions of the final foam. The foamable precursor into the mould is heated at temperatures above its melting point (Fig. 12 b). The material expands and fills the entire mould cavity. After that, the mould is cooled, followed by the extraction of the foam (Fig. 12c). The furnaces used to manufacture the metal foams usually are of the batch chamber furnace type (Fig. 13a). This research team has

(a) (b) Fig. 11. Al-alloy foam covered by a dense skin (a), with a closed-cell internal structure (b).

This method enables the cost effective production of aluminium alloy parts without limitations concerning shape ( e.g. panels, profiles or complex 3D shaped parts) (Duarte et al, 2006, 2008, 2010). Here, the group research results are presented to illustrate some practical examples (Fig. 14) which can be obtained by using the continuous foaming furnace

developed a foaming continuous furnace to produce these materials (Fig. 13b).

Fig. 10. Effect of the hot pressing temperature on the precursor density (a) and on the

00:00 00:10 00:20 00:30 00:40 00:50 01:00

Expansion [mm]

 500 °C 450°C 400°C 300 °C 200 °C

Time [hh:mm]

150 200 250 300 350 400 450 500 550 600

hot pressing temperature (ºC)

foaming behaviour (b).

**2.2 Production of metal foams** 

provide good surface finish (Fig. 11).

Percursor density (g/cm3

)

developed (Fig. 13 b). The disadvantage of the process is to produce components with a large volume parts as it is the case of Shinko Wire method (Miyoshi et al, 2000).

Fig. 12. Practical aspects of the process. (a) closed mould. (b) mould with the precursor into the pre-heated furnace. (c) Al-alloy foam block.

Fig. 13. Foaming furnaces. Batch chamber furnace (a). Continuous foaming furnace developed by this research team (b).

Fig. 14. 3D-parts of Al-alloy foams.

The atmosphere, heating rate and temperature of the thermal foaming cycle are some of the manufacturing parameters which influence the quality and the properties of the resulting foam. Moreover, it should also be referred that the characteristics of the mould (material, design and dimensions), the type of the furnace (batch or continuous) should also be considered (Duarte, 2005). The effects of these variables on the foaming behaviour may therefore be investigated. To assess the latter, foaming tests are performed by heating the

Aluminium Alloy Foams: Production and Properties 59

The final quality of the metal foams is evaluated by characterising its properties, namely the density, structural and mechanical properties. The foam density is relatively predictable and controlled by manipulation of the manufacturing parameters (Duarte & Banhart, 2005). Higher heating rates lead to an earlier expansion of the foamable precursor because the melting temperature is reached at an earlier time (Fig. 16). Besides that, the three expansion curves for the highest heating rates are quite similar. Only significantly lower heating rates lead to a change of the expansion characteristics, namely a lower maximum expansion. The reason for this may be the gas losses due to diffusion of hydrogen and perhaps the strong sample oxidation which might hinder expansion. In general, higher heating rate of the

Although at an industrial stage, the process still has major limitations on the ability to obtain tailor made cellular structure foams and to predict their properties. Cellular structured foams with different sizes and shapes of pores, and structural defects, can present a high density gradient depending on the component size (Fig. 17). These imperfections arise from the difficulties in controlling the manufacturing process such as: (i) lack of homogeneity in the precursor mixture (metal+blowing agent). (ii) lack of coordination between the mechanisms of metal melting and thermal decomposition of the blowing agent. (iii) difficulty in controlling the nucleation, growth and collapse stages. (iv) difficulty in

Fig. 17. Internal cellular structures for closed-cell AlSi7-alloy foam with different size.

The study of the physics of metal foaming is necessary to understand the underlying principle of metal foam formation and stabilization in order to produce better foams. This knowledge should lead to a control of these mechanisms during the foam formation. Currently, there are no detailed studies on the quantification of these mechanisms, only some aspects of foam evolution could be described by theoretical approaches. Despite the longstanding interest and research efforts regarding these mechanisms, the simulation and prediction of bubble nucleation in metallic foaming remains challenging. This is due to the difficulties in observing these mechanisms under experimental or actual processing conditions. For that, the researchers have been trying to quantify the involved mechanisms. Several investigations on foam formation have been carried out. The experimental techniques used for investigating the metallic foaming are mainly of two types: *ex-situ* techniques (Babcsán et al, 2005) and *in-situ* techniques (Babcsán et al, 2007). In the *ex-situ* techniques, the foaming process is interrupted by cooling at different foaming stages and

foamable precursor leads to the formation of foams with lower density.

stabilizing the foam formed and preventing the cells' collapse.

**3. The physics of metal foaming** 

precursor material up to its melting point inside of the apparatus called laser expandometer, which was specially developed and constructed for this purpose (Fig. 15). Here, the expansion (volume) and its temperature are monitored by means of a laser sensor and thermocouple, respectively. The measurement of the volume of the expanding melt together with the sample temperature generates a pair of functions V(t) and T(t) which characterise the foaming kinetics. The foaming process is strongly governed by temperature effects (Duarte & Banhart, 2000). The expansion curve is strongly dependent on the processing conditions, mainly on the hot pressing temperature to obtain the foamable precursor material, and the heating parameters during foaming (Fig. 10 b and Fig. 16).

Fig. 15. Laser expandometer used to characterise the foaming kinetics.

Fig. 16. Expansion curves of 6061 samples containing 0.6 wt. % TiH2, prepared with different heating rates (Duarte & Banhart, 2000).

precursor material up to its melting point inside of the apparatus called laser expandometer, which was specially developed and constructed for this purpose (Fig. 15). Here, the expansion (volume) and its temperature are monitored by means of a laser sensor and thermocouple, respectively. The measurement of the volume of the expanding melt together with the sample temperature generates a pair of functions V(t) and T(t) which characterise the foaming kinetics. The foaming process is strongly governed by temperature effects (Duarte & Banhart, 2000). The expansion curve is strongly dependent on the processing conditions, mainly on the hot pressing temperature to obtain the foamable precursor

material, and the heating parameters during foaming (Fig. 10 b and Fig. 16).

Fig. 15. Laser expandometer used to characterise the foaming kinetics.

120°C/mi 99°C/min

0

heating rates (Duarte & Banhart, 2000).

10

20

30

Expansion [mm]

40

50

60

00:00 00:10 00:20 00:30 01:00 01:30 02:00 02:30

86°C/min

6°C/min

Time [hh:mm]

Fig. 16. Expansion curves of 6061 samples containing 0.6 wt. % TiH2, prepared with different

The final quality of the metal foams is evaluated by characterising its properties, namely the density, structural and mechanical properties. The foam density is relatively predictable and controlled by manipulation of the manufacturing parameters (Duarte & Banhart, 2005). Higher heating rates lead to an earlier expansion of the foamable precursor because the melting temperature is reached at an earlier time (Fig. 16). Besides that, the three expansion curves for the highest heating rates are quite similar. Only significantly lower heating rates lead to a change of the expansion characteristics, namely a lower maximum expansion. The reason for this may be the gas losses due to diffusion of hydrogen and perhaps the strong sample oxidation which might hinder expansion. In general, higher heating rate of the foamable precursor leads to the formation of foams with lower density.

Although at an industrial stage, the process still has major limitations on the ability to obtain tailor made cellular structure foams and to predict their properties. Cellular structured foams with different sizes and shapes of pores, and structural defects, can present a high density gradient depending on the component size (Fig. 17). These imperfections arise from the difficulties in controlling the manufacturing process such as: (i) lack of homogeneity in the precursor mixture (metal+blowing agent). (ii) lack of coordination between the mechanisms of metal melting and thermal decomposition of the blowing agent. (iii) difficulty in controlling the nucleation, growth and collapse stages. (iv) difficulty in stabilizing the foam formed and preventing the cells' collapse.

Fig. 17. Internal cellular structures for closed-cell AlSi7-alloy foam with different size.

#### **3. The physics of metal foaming**

The study of the physics of metal foaming is necessary to understand the underlying principle of metal foam formation and stabilization in order to produce better foams. This knowledge should lead to a control of these mechanisms during the foam formation. Currently, there are no detailed studies on the quantification of these mechanisms, only some aspects of foam evolution could be described by theoretical approaches. Despite the longstanding interest and research efforts regarding these mechanisms, the simulation and prediction of bubble nucleation in metallic foaming remains challenging. This is due to the difficulties in observing these mechanisms under experimental or actual processing conditions. For that, the researchers have been trying to quantify the involved mechanisms. Several investigations on foam formation have been carried out. The experimental techniques used for investigating the metallic foaming are mainly of two types: *ex-situ* techniques (Babcsán et al, 2005) and *in-situ* techniques (Babcsán et al, 2007). In the *ex-situ* techniques, the foaming process is interrupted by cooling at different foaming stages and

Aluminium Alloy Foams: Production and Properties 61

The foaming process can be divided into three stages: bubble nucleation, bubble growth and foam collapse (Fig. 18b). Duarte and Banhart have done some pioneer research, in the understanding of the mechanisms involved in the PM process, which contributed to the significant advance of the state of the art (Duarte & Banhart, 2000), using *ex-situ* technique in which the foaming process is interrupted by cooling at different foaming stages and the resulting solid foam is characterized. Topics such as bubble nucleation, bubble growth and foam stability were discussed. As a result of this research, it is concluded that the bubbles nucleation occurs usually in the solid state, the pressure generated by the gas can deform the metallic matrix, the process is governed by the principle of semisolid metal processing, the bubble growth dynamics is controlled by the thermal decomposition of the blowing

00:00 00:05 00:10 00:45 00:50 00:55 01:00

Time (hh:mm) (a)

(b)

**A B C D** Drainage effects **E** 

Fig. 18. (a) Expansion curve of precursor containing AA 6061 sample containing 0.6 wt. % TiH2 foamed in a pre-heated furnace at 800°C. (b) Morphology of the AA 6061 foams in different foaming stages (foam diameter 30 mm). The letters show at which expansion

Bubble nucleation Bubble growth Foam collapse

C

B A

D E

Coalescence effects

agent and the metal melting.

0

10

20

30

Expansion (mm)

Foaming direction

stage the sample was removed (see Fig. 15).

40

50

60

the resulting solid foam is characterized. The disadvantage of this approach is that it takes a long time to carry out such investigation, and that the results suffer from a certain inaccuracy originating statistical variations between a single experiment. Even if the starting materials for the individual foaming tests were produced in the same way, each foaming experimental test would turn out slightly differently, due to effects such as agglomerates of the blowing agent particles, structural defects and impurities in the precursor (Duarte, 2005). The 3D-image of X-ray tomographic observations of solid foam samples in different foaming stages allow to observe the modification of the shape of the bubbles (Stanzick et al, 2002).

In the *in-situ* techniques, the foaming studies are evaluated during the evolution of one single sample (Garcia-Moreno et al, 2005; Rack et al, 2009). Neutron radioscopy and X-ray radioscopy were employed for *in situ* observation drainage mechanisms, the early stages and growth stages, during the foaming process (Bellman et al, 2002). The temporal development of the cellular structures of liquid metallic foams and the redistributions of the metal can be observed by synchrotron based on neutron radioscopy.

Theoretical studies of the foaming process itself concentrate predominantly on the detailed analysis of microscopic evolution of foams, mostly on the basis of an already existing cellular structure (Stavans, 1993). A theoretical study on metal foam processing has treated the material flow behaviour and focused on the description of the solidification stage during the process using an one-dimensional model which combines the equations of foam drainage with Fourier equation. However, this model is not able to describe the entire foaming process starting from bubble nucleation to final foam development. Other authors showed that the standard foam drainage equation (FDE) can principally be used to describe drainage in metal foams (Körner et al, 2008; Brunke & Odenbach, 2006; Belkessam & Fristching, 2003). However, the effective viscosity was considered to be one order of magnitude higher than the original one to match the experimental observations (Stanzick et al, 2002). A published model of metal foaming based on a Lattice–Boltzmann procedure (Körner et al, 2002) treats the foaming problem in more detail, (i.e. from bubble nucleation to the resulting foam structure, however, without considering the chemical decomposition of the blowing agent as well as the thermal heating process). The foam is considered in the liquid state; i.e, melting and solidification are not taken into account.

These studies have had significant roles in contributing to a more complete understanding of bubble-growth phenomena. However, in studying bubble-growth behaviour, almost all of these previous works involve pure theoretical studies without experimental verification, since only limited experiments have addressed the dynamic behaviour of the phenomena. Moreover, some of the physical parameters that were used to describe the materials adopted in these theoretical studies were unrealistic. There are no mathematical models available for foam evolution including nucleation, growth, coalescence and decay. Generally, numerical or analytical models focus on a particular phenomenon (e.g. drainage). The analytical and numerical approaches available in the literature are very limited. From the engineering point of view it is a very complex process because it involves several physical, chemical, thermal and mechanical phenomena that occur at the same time or successively.

#### **3.1 Foam evolution**

The evolution of the cellular pores during the foaming process has been evaluated and discussed using the *ex-situ* and *in-situ* techniques, as shown in the presented examples in Fig. 18 and 19, respectively.

the resulting solid foam is characterized. The disadvantage of this approach is that it takes a long time to carry out such investigation, and that the results suffer from a certain inaccuracy originating statistical variations between a single experiment. Even if the starting materials for the individual foaming tests were produced in the same way, each foaming experimental test would turn out slightly differently, due to effects such as agglomerates of the blowing agent particles, structural defects and impurities in the precursor (Duarte, 2005). The 3D-image of X-ray tomographic observations of solid foam samples in different foaming stages allow to observe the modification of the shape of the bubbles (Stanzick et al, 2002). In the *in-situ* techniques, the foaming studies are evaluated during the evolution of one single sample (Garcia-Moreno et al, 2005; Rack et al, 2009). Neutron radioscopy and X-ray radioscopy were employed for *in situ* observation drainage mechanisms, the early stages and growth stages, during the foaming process (Bellman et al, 2002). The temporal development of the cellular structures of liquid metallic foams and the redistributions of the

Theoretical studies of the foaming process itself concentrate predominantly on the detailed analysis of microscopic evolution of foams, mostly on the basis of an already existing cellular structure (Stavans, 1993). A theoretical study on metal foam processing has treated the material flow behaviour and focused on the description of the solidification stage during the process using an one-dimensional model which combines the equations of foam drainage with Fourier equation. However, this model is not able to describe the entire foaming process starting from bubble nucleation to final foam development. Other authors showed that the standard foam drainage equation (FDE) can principally be used to describe drainage in metal foams (Körner et al, 2008; Brunke & Odenbach, 2006; Belkessam & Fristching, 2003). However, the effective viscosity was considered to be one order of magnitude higher than the original one to match the experimental observations (Stanzick et al, 2002). A published model of metal foaming based on a Lattice–Boltzmann procedure (Körner et al, 2002) treats the foaming problem in more detail, (i.e. from bubble nucleation to the resulting foam structure, however, without considering the chemical decomposition of the blowing agent as well as the thermal heating process). The foam is considered in the

These studies have had significant roles in contributing to a more complete understanding of bubble-growth phenomena. However, in studying bubble-growth behaviour, almost all of these previous works involve pure theoretical studies without experimental verification, since only limited experiments have addressed the dynamic behaviour of the phenomena. Moreover, some of the physical parameters that were used to describe the materials adopted in these theoretical studies were unrealistic. There are no mathematical models available for foam evolution including nucleation, growth, coalescence and decay. Generally, numerical or analytical models focus on a particular phenomenon (e.g. drainage). The analytical and numerical approaches available in the literature are very limited. From the engineering point of view it is a very complex process because it involves several physical, chemical,

The evolution of the cellular pores during the foaming process has been evaluated and discussed using the *ex-situ* and *in-situ* techniques, as shown in the presented examples in

thermal and mechanical phenomena that occur at the same time or successively.

**3.1 Foam evolution** 

Fig. 18 and 19, respectively.

metal can be observed by synchrotron based on neutron radioscopy.

liquid state; i.e, melting and solidification are not taken into account.

The foaming process can be divided into three stages: bubble nucleation, bubble growth and foam collapse (Fig. 18b). Duarte and Banhart have done some pioneer research, in the understanding of the mechanisms involved in the PM process, which contributed to the significant advance of the state of the art (Duarte & Banhart, 2000), using *ex-situ* technique in which the foaming process is interrupted by cooling at different foaming stages and the resulting solid foam is characterized. Topics such as bubble nucleation, bubble growth and foam stability were discussed. As a result of this research, it is concluded that the bubbles nucleation occurs usually in the solid state, the pressure generated by the gas can deform the metallic matrix, the process is governed by the principle of semisolid metal processing, the bubble growth dynamics is controlled by the thermal decomposition of the blowing agent and the metal melting.

Fig. 18. (a) Expansion curve of precursor containing AA 6061 sample containing 0.6 wt. % TiH2 foamed in a pre-heated furnace at 800°C. (b) Morphology of the AA 6061 foams in different foaming stages (foam diameter 30 mm). The letters show at which expansion stage the sample was removed (see Fig. 15).

Aluminium Alloy Foams: Production and Properties 63

the system (metal+gas). The metallic foaming rheology is not simple and its mechanical,

The fundamental stability, collapse and solidification mechanisms during the foam formation have been investigated. These topics are the most controversial ones in the metal foam research. A better understanding of these mechanisms is required for an accurate control of the foam structures, such as cell size and porosity. *Ex-situ* and *in-situ* techniques have been used in this field. *In-situ* techniques have been used to measure the expansion, the density evolution and the effects of the drainage and coalescence. The effects of the thermal foaming cycle (heating rate, temperature and atmosphere) on the foam stability and foam collapse parameters have been also studied and observed (Banhart, 2006). PM foams belong to the class of transient (unstable) foams with lifetimes of seconds or to the permanent (metastable) foams with lifetimes of hours. The foamability is thought to result from the Gibbs–Marangoni effect where a membrane is stabilized during thinning due to liquid metal flow towards the weakened region, because of a local increase in surface tension. The flow is the response to a surface tension gradient. Due to viscous drag the flow can carry an appreciable amount of underlying liquid along with it so that it restores the thickness. In contrast to transient foams, the film thinning times in permanent foams are relatively short compared to the lifetime. The stability is controlled by the balance of interfacial forces. These forces equilibrate after drainage has been completed. The effect of temperature and gravity have been evaluated via observation of bubble size, ruptures of the bubbles, relative distribution of ruptures between bottom and top, draining and timing of draining, as well as

Foam decays by combination of three phenomena – gas loss, drainage and coalescence (Duarte, 2005). Part of the gas is lost by diffusion from the outer surface to the surrounding. In the liquid state, loss is expected to rise due to a higher diffusivity than in its solid state. However, crack formation during expansion at the outer surface of precursors can also lead to gas losses. Moreover, bubble rupture at outer surface also results in sudden gas loss. The second and third way of gas loss are random events and not much effort was attributed to asses their contribution except for some qualitative statement. Drainage is one of the driving forces for the temporal instability of liquid foams, caused mainly by gravitational and capillary effects. The drainage and coalescence mechanisms can be observed using *ex-situ* techniques with the formation of a thick metal layer on the bottom of the solid foam and large pores, respectively. Foam shows a very complex rheological behaviour including the

Foam solidification is also an essential processing step of foam production. An uncontrolled solidification can create defects (e.g. cracks) in the cell wall of the foams (Duarte, 2005). So, this study helps to reduce the defects observed in final structures and thereby to improve the mechanical performance of the foam. The solidification by cooling of a foamed liquid metal is a "race against time", in as much as the relatively heavy liquid is prone to drainage, which rapidly reduces the foam density and hence provokes instability and collapse. The foamed liquid is immediately subjected to gravity-driven drainage of liquid, creating a vertical profile of density (or liquid fraction). At any point in the sample this adjustment must proceed until the freezing point is reached. Thus, at intermediate times, the sample

thermal and chemical interaction leads to coupled problems of great complexity.

**3.2 Foam: Collapse, stabilisation and solidification**

the effect of temperature distribution.

bubble deformation, rearrangement and avalanches processes.

Fig. 19. In-situ radioscopic images acquired during a foaming process of a precursor.

The foaming process can be divided into three stages: bubble nucleation, bubble growth and foam collapse (Fig. 18). The shape of the bubbles varies during the foaming process. The bubbles appear as cracks aligned perpendicular to the foaming direction, changing to a spherical geometry, followed by polyhedral geometries. As described above, the PM method consists on heating a solid precursor material (Fig. 9). When heated, the metal melts (e.g. Al alloy) and the blowing agent (e.g. TiH2) produces a gas that creates the bubbles in the foam. The heating process leads to partial metal melting as well as to the release of the gas and consequently to the material foaming in its semi-solid state of the material. The heat supply takes place in the solid material up to the *solidus* temperature. The decomposition of the blowing agent may start in the still with the precursor in solid state, so that the bubble nucleation can be previously initiated here. For this reason, the pores formed at early stages of foaming appear as cracks aligned perpendicular to the foaming direction (Fig. 18b in the nucleation stage). Moreover, the partial decomposition of the blowing agent particles during the compaction step can occur in which the temperature in this step is near to the initial decomposition temperature. The released gas in this compaction step is entrapping into the matrix metal, and can create a sufficient number of initial nuclei into the foamable precursor in which can act as centre of bubble nucleation. Gas accumulates in residual porosity and builds up pressure as temperature increases. With the occurrence of bubble growth, a twophase flow (liquid metal and gas bubbles) develops above the *solidus* temperature.

After the alloy melting, the crack-like pores round off to minimize surface energy. Bubble growth begins, driven by gas release from the blowing agent, and the structure starts to appear as foam. With the increase of the temperature the internal gas pressure of each nucleated bubble increase, and turn in the strength of the metal matrix is reduced down to its value at the melting point. The bubble growth may not be uniform because depends on the characteristics of the foamable precursor material. The elongated initial bubbles are increasing in size and become more spherical (Fig. 18). The more spherical bubbles are observed in the cellular structures when the expansion reaches the maximum value. The spherical bubbles become more polyhedrical bubbles. After maximum expansion, no more gas is released and the foam begins to collapse. The latter is due the drainage and coalescence mechanisms, which are discussed ahead.

Foam growth depends not only on the rheology of the system metal/gas and mechanical strength but also on the pressure inside each bubble and its architecture. This growth is affected by various factors such as the content and distribution of the blowing agent, the hydrostatic pressure or tension applied to the metallic matrix and the viscous properties of

Fig. 19. In-situ radioscopic images acquired during a foaming process of a precursor.

phase flow (liquid metal and gas bubbles) develops above the *solidus* temperature.

coalescence mechanisms, which are discussed ahead.

After the alloy melting, the crack-like pores round off to minimize surface energy. Bubble growth begins, driven by gas release from the blowing agent, and the structure starts to appear as foam. With the increase of the temperature the internal gas pressure of each nucleated bubble increase, and turn in the strength of the metal matrix is reduced down to its value at the melting point. The bubble growth may not be uniform because depends on the characteristics of the foamable precursor material. The elongated initial bubbles are increasing in size and become more spherical (Fig. 18). The more spherical bubbles are observed in the cellular structures when the expansion reaches the maximum value. The spherical bubbles become more polyhedrical bubbles. After maximum expansion, no more gas is released and the foam begins to collapse. The latter is due the drainage and

Foam growth depends not only on the rheology of the system metal/gas and mechanical strength but also on the pressure inside each bubble and its architecture. This growth is affected by various factors such as the content and distribution of the blowing agent, the hydrostatic pressure or tension applied to the metallic matrix and the viscous properties of

The foaming process can be divided into three stages: bubble nucleation, bubble growth and foam collapse (Fig. 18). The shape of the bubbles varies during the foaming process. The bubbles appear as cracks aligned perpendicular to the foaming direction, changing to a spherical geometry, followed by polyhedral geometries. As described above, the PM method consists on heating a solid precursor material (Fig. 9). When heated, the metal melts (e.g. Al alloy) and the blowing agent (e.g. TiH2) produces a gas that creates the bubbles in the foam. The heating process leads to partial metal melting as well as to the release of the gas and consequently to the material foaming in its semi-solid state of the material. The heat supply takes place in the solid material up to the *solidus* temperature. The decomposition of the blowing agent may start in the still with the precursor in solid state, so that the bubble nucleation can be previously initiated here. For this reason, the pores formed at early stages of foaming appear as cracks aligned perpendicular to the foaming direction (Fig. 18b in the nucleation stage). Moreover, the partial decomposition of the blowing agent particles during the compaction step can occur in which the temperature in this step is near to the initial decomposition temperature. The released gas in this compaction step is entrapping into the matrix metal, and can create a sufficient number of initial nuclei into the foamable precursor in which can act as centre of bubble nucleation. Gas accumulates in residual porosity and builds up pressure as temperature increases. With the occurrence of bubble growth, a two-

t1 t2 t3 t4 t5 t6 Bubble nucleation Bubble growth

the system (metal+gas). The metallic foaming rheology is not simple and its mechanical, thermal and chemical interaction leads to coupled problems of great complexity.

#### **3.2 Foam: Collapse, stabilisation and solidification**

The fundamental stability, collapse and solidification mechanisms during the foam formation have been investigated. These topics are the most controversial ones in the metal foam research. A better understanding of these mechanisms is required for an accurate control of the foam structures, such as cell size and porosity. *Ex-situ* and *in-situ* techniques have been used in this field. *In-situ* techniques have been used to measure the expansion, the density evolution and the effects of the drainage and coalescence. The effects of the thermal foaming cycle (heating rate, temperature and atmosphere) on the foam stability and foam collapse parameters have been also studied and observed (Banhart, 2006). PM foams belong to the class of transient (unstable) foams with lifetimes of seconds or to the permanent (metastable) foams with lifetimes of hours. The foamability is thought to result from the Gibbs–Marangoni effect where a membrane is stabilized during thinning due to liquid metal flow towards the weakened region, because of a local increase in surface tension. The flow is the response to a surface tension gradient. Due to viscous drag the flow can carry an appreciable amount of underlying liquid along with it so that it restores the thickness. In contrast to transient foams, the film thinning times in permanent foams are relatively short compared to the lifetime. The stability is controlled by the balance of interfacial forces. These forces equilibrate after drainage has been completed. The effect of temperature and gravity have been evaluated via observation of bubble size, ruptures of the bubbles, relative distribution of ruptures between bottom and top, draining and timing of draining, as well as the effect of temperature distribution.

Foam decays by combination of three phenomena – gas loss, drainage and coalescence (Duarte, 2005). Part of the gas is lost by diffusion from the outer surface to the surrounding. In the liquid state, loss is expected to rise due to a higher diffusivity than in its solid state. However, crack formation during expansion at the outer surface of precursors can also lead to gas losses. Moreover, bubble rupture at outer surface also results in sudden gas loss. The second and third way of gas loss are random events and not much effort was attributed to asses their contribution except for some qualitative statement. Drainage is one of the driving forces for the temporal instability of liquid foams, caused mainly by gravitational and capillary effects. The drainage and coalescence mechanisms can be observed using *ex-situ* techniques with the formation of a thick metal layer on the bottom of the solid foam and large pores, respectively. Foam shows a very complex rheological behaviour including the bubble deformation, rearrangement and avalanches processes.

Foam solidification is also an essential processing step of foam production. An uncontrolled solidification can create defects (e.g. cracks) in the cell wall of the foams (Duarte, 2005). So, this study helps to reduce the defects observed in final structures and thereby to improve the mechanical performance of the foam. The solidification by cooling of a foamed liquid metal is a "race against time", in as much as the relatively heavy liquid is prone to drainage, which rapidly reduces the foam density and hence provokes instability and collapse. The foamed liquid is immediately subjected to gravity-driven drainage of liquid, creating a vertical profile of density (or liquid fraction). At any point in the sample this adjustment must proceed until the freezing point is reached. Thus, at intermediate times, the sample

Aluminium Alloy Foams: Production and Properties 65

deformation energy absorbed under compression (Markaki & Clyne, 2001; Campana & Pilone, 2008). Mechanical studies demonstrate that selective deformation of the weakest region of the foam structure leads to crush-band formation (Duarte et al, 2009). Cell morphology and interconnection could also affect thermal and acoustic properties (Kolluri et al, 2008). It is widely accepted that foams with a uniform pore distribution and defects free, are desirable. This would make the properties more predictable. Only then, metal foams will be considered reliable materials for engineering purposes and will be able to compete with classical materials. Despite their quality improvement in the last 10 years the resulting metal foams still suffer from non-uniformities. Scientists aim to produce more regular structures with fewer defects in a more reproducible way which is the crucial challenge of the research in this field. Foam characterisation results revealed that the cellular structures of the Al-alloy foams obtained by PM method have pores with different sizes and shapes (Fig. 20). A large size distribution of the cellular pores with irregular cell shape is observed. The closed pores aremostly- of polyhedral or spherical geometry (Fig. 20 and 21). Spherical pores with a thick thickness of cell-wall are mostly observed in the bottom and lateral sides of the foam samples (Figs. 20 and 21a). Polyhedral pores with a thin cell-wall thickness are mainly observed at the top of the foam samples (Figs. 20 and 21b). The distribution of the solid metal in the foam is also non-uniform and leads to a higher density gradient (Fig.20). These materials have a broader cell diameter distribution curve (Fig. 20c). The cell-size distribution is dominated by high number of the small pores. The most of the pores have diameter lower than 2mm. The magnified images of the cross section of the sample reveal small porosities in the dense surface skin (Fig. 22). Significant morphological defects such as cracks or spherical micropores in cell walls and cell wall wiggles and dense surface skin are also observed. Each cell has normally approximately 5 other ones in its vicinity (Fig. 21). The distribution of the cell-wall thickness has an asymmetric shape for these foams (Fig.21). The smallest cell-wall thickness is about 70 m. The maximum cell-wall thickness is about 500 m. The thickness of the cell wall depends on the foam density. The thickness of the external dense surface skin around sample varies, where the higher values are located in the lateral and bottom sides. AlSi7 foams presents 565.56 m, 365.40 m and 214.58 m, respectively for bottom, lateral and top sides of the samples (Fig. 22). Other structural feature that affects the mechanical behaviour is the microstructure of the massive cell material. Depending on the

<2 mm 2-4 mm 4-6mm >6mm

 AlSi7-foams AlSi1Mg-foams

Pore range [mm]

Pores number

(a) (b) (c) Fig. 20. Cellular structures of AlSi7 (a) and AlSi1Mg (b) foams. Cell-pore size distribution as

a function of the number of cell pores for both foams (c).

consists of a solidified outer shell surrounding a draining liquid core. The competition between drainage and heat transfer, leading to solidification was studied by Mukherjee *et al* using X-ray radioscopy (Mukherjee et al, 2009). A hitherto unknown expansion stage was observed during solidification of Al-alloy foams. The phenomena that occur simultaneously while foam solidifies are associated to its volume change (Mukherjee et al, 2009). The extra expansion is observed and it is an anomalous behaviour since the foam is expected to contract during solidification. This extra expansion takes place whenever the combined volume gain rate is more than the combined volume loss rate. Moreover, it increases as the cooling rate decreases. Mukherjee reported that the slow cooling of foams can trigger extra expansion which in turn can induce defects in foam morphology. This is due to the extra expansion induced rupture during solidification of the metallic melt inside the cell wall.

#### **4. Properties**

The properties of the metal foams belong to a group of materials called cellular solids which are defined as having porosity >up to 0.7 (Gibson & Ashby, 1997). Natural foams are produced by plants and animals such as cork or bone. Man made foams can be manufactured from a variety of materials such as ceramics, polymers and metals. There are two categories of foams: open - and closed- cells. Here, it is presented a brief overview of the main properties of closed-cell Al-alloy foams obtained by the PM method.

Metal foams combine properties of cellular materials with those of metals. For this reason, metal foams are advantageous for lightweight constructions due to their high strength-toweight ratio, in combination with structural and functional properties like crash energy absorption, sound and heat management (Asbhy et al, 2000; Degischer & Kriszt, 2002). Many metals and their alloys can be foamed. Among the metal foams, the Al-alloy ones are commercially the most exploited due to their low density, high ductility, high thermal conductivity, and metal competitive cost.

#### **4.1 Structural properties**

There are several structural parameters of these foams, such as number, size-pore distribution, average size, shape and geometry of the pores, thickness, intersections and defects in the cell-walls and thickness, defects and cracks of the external dense surface for describing the cellular architecture of the foams. The properties of these foams are influenced by these morphological features (Gibson &Ashby, 1997; Ramamurty & Paul, 2004; Campana et al, 2008).

Progress has been made in understanding the relationship between properties and morphology. Although this exact interrelationship is not yet sufficiently known, one usually assumes that the properties are improved when all the individual cells of a foam have similar size and a spherical shape. This has not really been verified experimentally. There is no doubt that the density of a metal foam and the matrix alloy properties influence the modulus and strength of the foam. All studies indicate that the real properties are inferior than the theoretically expected due to structural defects. This demands a better pore control and reduction in structural defects. Density variation and imperfections yield a large scatter of measured properties, which is detrimental for the metal foams reliability (Ramamurty & Paul, 2004). Wiggled or missing cell-walls reduce strength, and in turn, result in a reduced

consists of a solidified outer shell surrounding a draining liquid core. The competition between drainage and heat transfer, leading to solidification was studied by Mukherjee *et al* using X-ray radioscopy (Mukherjee et al, 2009). A hitherto unknown expansion stage was observed during solidification of Al-alloy foams. The phenomena that occur simultaneously while foam solidifies are associated to its volume change (Mukherjee et al, 2009). The extra expansion is observed and it is an anomalous behaviour since the foam is expected to contract during solidification. This extra expansion takes place whenever the combined volume gain rate is more than the combined volume loss rate. Moreover, it increases as the cooling rate decreases. Mukherjee reported that the slow cooling of foams can trigger extra expansion which in turn can induce defects in foam morphology. This is due to the extra expansion induced rupture during solidification of the metallic melt inside the cell wall.

The properties of the metal foams belong to a group of materials called cellular solids which are defined as having porosity >up to 0.7 (Gibson & Ashby, 1997). Natural foams are produced by plants and animals such as cork or bone. Man made foams can be manufactured from a variety of materials such as ceramics, polymers and metals. There are two categories of foams: open - and closed- cells. Here, it is presented a brief overview of the

Metal foams combine properties of cellular materials with those of metals. For this reason, metal foams are advantageous for lightweight constructions due to their high strength-toweight ratio, in combination with structural and functional properties like crash energy absorption, sound and heat management (Asbhy et al, 2000; Degischer & Kriszt, 2002). Many metals and their alloys can be foamed. Among the metal foams, the Al-alloy ones are commercially the most exploited due to their low density, high ductility, high thermal

There are several structural parameters of these foams, such as number, size-pore distribution, average size, shape and geometry of the pores, thickness, intersections and defects in the cell-walls and thickness, defects and cracks of the external dense surface for describing the cellular architecture of the foams. The properties of these foams are influenced by these morphological features (Gibson &Ashby, 1997; Ramamurty & Paul,

Progress has been made in understanding the relationship between properties and morphology. Although this exact interrelationship is not yet sufficiently known, one usually assumes that the properties are improved when all the individual cells of a foam have similar size and a spherical shape. This has not really been verified experimentally. There is no doubt that the density of a metal foam and the matrix alloy properties influence the modulus and strength of the foam. All studies indicate that the real properties are inferior than the theoretically expected due to structural defects. This demands a better pore control and reduction in structural defects. Density variation and imperfections yield a large scatter of measured properties, which is detrimental for the metal foams reliability (Ramamurty & Paul, 2004). Wiggled or missing cell-walls reduce strength, and in turn, result in a reduced

main properties of closed-cell Al-alloy foams obtained by the PM method.

conductivity, and metal competitive cost.

**4.1 Structural properties** 

2004; Campana et al, 2008).

**4. Properties** 

deformation energy absorbed under compression (Markaki & Clyne, 2001; Campana & Pilone, 2008). Mechanical studies demonstrate that selective deformation of the weakest region of the foam structure leads to crush-band formation (Duarte et al, 2009). Cell morphology and interconnection could also affect thermal and acoustic properties (Kolluri et al, 2008). It is widely accepted that foams with a uniform pore distribution and defects free, are desirable. This would make the properties more predictable. Only then, metal foams will be considered reliable materials for engineering purposes and will be able to compete with classical materials. Despite their quality improvement in the last 10 years the resulting metal foams still suffer from non-uniformities. Scientists aim to produce more regular structures with fewer defects in a more reproducible way which is the crucial challenge of the research in this field.

Foam characterisation results revealed that the cellular structures of the Al-alloy foams obtained by PM method have pores with different sizes and shapes (Fig. 20). A large size distribution of the cellular pores with irregular cell shape is observed. The closed pores aremostly- of polyhedral or spherical geometry (Fig. 20 and 21). Spherical pores with a thick thickness of cell-wall are mostly observed in the bottom and lateral sides of the foam samples (Figs. 20 and 21a). Polyhedral pores with a thin cell-wall thickness are mainly observed at the top of the foam samples (Figs. 20 and 21b). The distribution of the solid metal in the foam is also non-uniform and leads to a higher density gradient (Fig.20). These materials have a broader cell diameter distribution curve (Fig. 20c). The cell-size distribution is dominated by high number of the small pores. The most of the pores have diameter lower than 2mm. The magnified images of the cross section of the sample reveal small porosities in the dense surface skin (Fig. 22). Significant morphological defects such as cracks or spherical micropores in cell walls and cell wall wiggles and dense surface skin are also observed. Each cell has normally approximately 5 other ones in its vicinity (Fig. 21). The distribution of the cell-wall thickness has an asymmetric shape for these foams (Fig.21). The smallest cell-wall thickness is about 70 m. The maximum cell-wall thickness is about 500 m. The thickness of the cell wall depends on the foam density. The thickness of the external dense surface skin around sample varies, where the higher values are located in the lateral and bottom sides. AlSi7 foams presents 565.56 m, 365.40 m and 214.58 m, respectively for bottom, lateral and top sides of the samples (Fig. 22). Other structural feature that affects the mechanical behaviour is the microstructure of the massive cell material. Depending on the

Fig. 20. Cellular structures of AlSi7 (a) and AlSi1Mg (b) foams. Cell-pore size distribution as a function of the number of cell pores for both foams (c).

Aluminium Alloy Foams: Production and Properties 67

of these foams on the mechanical properties are strong and complex. Depending on the material from which the foam is made, different mechanisms (brittle or ductile) can be observed. The compressive properties, such as, the average plateau stress, modulus, the elasticity and the energy absorption, depend, above all, on the foam density in which their

High density High density

Fig. 23. Compressive stress-strain curves for the different specimens (cylindrical samples

As it can be depicted from Fig. 23, the stress-strain curves are divided into three characteristic regions. The first region (I) is linear-elastic where the load increases with increasing compression displacement almost linearly (elastic deflection of the pore walls), followed by a plastic collapse plateau (Region II) with a nearly steady compression load (pore walls yield or fracture, whereas increasing deformation does not require an increase of the load). The last region is the densification of the foam (Region III) where there is a rapid

These foams exhibit, after an elastic loading, a more or less clear plateau region. This plateau stress is important to characterise the energy absorbing behaviour and is a good material property for the compression performance of a foam. The measurements of the plateau stress depending on the different methods exist for measurement of the plateau stress

The failure modes and mechanisms associated to these foams at different regions of the

The elastic deformation occurs due to bending of the edges, elongation of cell walls and trapped gas pressure inside the cells. The deformation is not visible (point B, in Fig.24b). The

0 10 20 30 40 50 60 70 80 90 100

Strain (%)

 0.38 g/cm3 0.39 g/cm3 0.42 g/cm3 0.44 g/cm3 0.51 g/cm3 0.52 g/cm3 0.55 g/cm3 0.56 g/cm3

AlSi1Mg foams

values increase with the rise of the density (Fig. 23).

AlSi7 foams 0.47 g/cm3 0.54 g/cm3 0.55 g/cm3 0.58 g/cm3 0.58 g/cm3 0.58g/cm3 0.60g/cm3

0 10 20 30 40 50 60 70 80 90 100

Strain (%)

height/diameter ratio equal 1: h== 30mm x 30mm).

increase on the load after the cell walls crushed together.

depending on the course of the stress-strain curve.

load-displacement have been identified (Fig. 24).

Stress (MPa)

alloy composition and on the manufacturing process, metallic dendrites, eutectic cells, precipitates, or even particles can be observed in the final cellular structures. Foams having same density but made of different Al-alloys can reached different plateau stress.

Fig. 21. Pore geometries: (a) spherical, (b) polyhedral and (c) other geometries.

Fig. 22. Thicknesses of the dense surface skin in different sections a sample AlSi7 foam: (a) top, (b) lateral and (b) bottom sides.

#### **4.2 Mechanical properties**

Many literature studies have been undertaken on the mechanical properties of metal foams. A broad survey of the understanding of the mechanical behaviour of a wide range of cellular solids is provided by Gibson and Ashby (Gibson & Ashby, 2000). Others, have carried out experiments to investigate the behaviour of metallic foams under different loading conditions, particularly the properties of metal foams under impact loading. The possibility of controlling the load-displacement behaviour by an appropriate selection of matrix material, cellular geometry and relative density makes foams an ideal material for energy absorbing structures. Among the several mechanical testing methods available, uniaxial compressive mechanical tests are commonly used to evaluate the compressive behaviour and the energy absorbed of these foams. The elastic modulus, yield and plateau strengths are the most important mechanical properties parameters which are obtained from these curves. The stress-strain curves of closed-cell Al-alloy foams display either plastic or brittle fracture depending on foam fabrication and microstructure (Sugimura et al, 1997; Banhart & Baumeister, 1998).

The compression behaviour of these Al-alloy foams depends on several parameters such as: (i) the Al-alloy composition; (ii) the foam morphology (cell size range); (iii) the density gradient of samples; (iv) the defects of cellular structure (cell walls) and (v) the characteristics of the external surface skin. The influence of the density and the architecture

alloy composition and on the manufacturing process, metallic dendrites, eutectic cells, precipitates, or even particles can be observed in the final cellular structures. Foams having

(a) (b) (c)

(a) (b) (c) Fig. 22. Thicknesses of the dense surface skin in different sections a sample AlSi7 foam: (a)

Many literature studies have been undertaken on the mechanical properties of metal foams. A broad survey of the understanding of the mechanical behaviour of a wide range of cellular solids is provided by Gibson and Ashby (Gibson & Ashby, 2000). Others, have carried out experiments to investigate the behaviour of metallic foams under different loading conditions, particularly the properties of metal foams under impact loading. The possibility of controlling the load-displacement behaviour by an appropriate selection of matrix material, cellular geometry and relative density makes foams an ideal material for energy absorbing structures. Among the several mechanical testing methods available, uniaxial compressive mechanical tests are commonly used to evaluate the compressive behaviour and the energy absorbed of these foams. The elastic modulus, yield and plateau strengths are the most important mechanical properties parameters which are obtained from these curves. The stress-strain curves of closed-cell Al-alloy foams display either plastic or brittle fracture depending on foam fabrication and microstructure (Sugimura et al, 1997;

The compression behaviour of these Al-alloy foams depends on several parameters such as: (i) the Al-alloy composition; (ii) the foam morphology (cell size range); (iii) the density gradient of samples; (iv) the defects of cellular structure (cell walls) and (v) the characteristics of the external surface skin. The influence of the density and the architecture

same density but made of different Al-alloys can reached different plateau stress.

Fig. 21. Pore geometries: (a) spherical, (b) polyhedral and (c) other geometries.

top, (b) lateral and (b) bottom sides.

**4.2 Mechanical properties** 

Banhart & Baumeister, 1998).

of these foams on the mechanical properties are strong and complex. Depending on the material from which the foam is made, different mechanisms (brittle or ductile) can be observed. The compressive properties, such as, the average plateau stress, modulus, the elasticity and the energy absorption, depend, above all, on the foam density in which their values increase with the rise of the density (Fig. 23).

Fig. 23. Compressive stress-strain curves for the different specimens (cylindrical samples height/diameter ratio equal 1: h== 30mm x 30mm).

As it can be depicted from Fig. 23, the stress-strain curves are divided into three characteristic regions. The first region (I) is linear-elastic where the load increases with increasing compression displacement almost linearly (elastic deflection of the pore walls), followed by a plastic collapse plateau (Region II) with a nearly steady compression load (pore walls yield or fracture, whereas increasing deformation does not require an increase of the load). The last region is the densification of the foam (Region III) where there is a rapid increase on the load after the cell walls crushed together.

These foams exhibit, after an elastic loading, a more or less clear plateau region. This plateau stress is important to characterise the energy absorbing behaviour and is a good material property for the compression performance of a foam. The measurements of the plateau stress depending on the different methods exist for measurement of the plateau stress depending on the course of the stress-strain curve.

The failure modes and mechanisms associated to these foams at different regions of the load-displacement have been identified (Fig. 24).

The elastic deformation occurs due to bending of the edges, elongation of cell walls and trapped gas pressure inside the cells. The deformation is not visible (point B, in Fig.24b). The

Aluminium Alloy Foams: Production and Properties 69

absorption capability of these foams can be well estimated from the stress-strain compression behaviour of the material which is estimated from the area under the stressstrain curve (Fig. 25a). As foam materials exhibit a constant stress "plateau" they can absorb higher levels of energy than dense aluminium alloys. Most of the absorbed energy is irreversibly converted into a plastic deformation energy which is a further advantage of foamed Aluminium. For the same stress level, the dense material is deformed in the regime of reversible linear-elastic stresses, releasing most of the stored energy when the load is removed. Al-alloy foams exhibit higher energy absorption capabilities (Fig. 25). The increase

0 20 40 60 80 100

Strain (%)

0

5

10

Energy (MJ/m3

(a) (b)

Despite its technological advances, the metallic foam formation is not problem free and still poses challenges. Questions related to a very hot topic (i.e. the control of the pores size and shape of metal foams) that is seen with alacrity by the scientific community due to potential applications of these materials in the transport industry, are highlighted and discussed in detail. The key-question is how to produce metal foams, in series, achieving uniform cellular structure, in order to improve the manufacture reproducibility and to control foam architecture. A key goal of this group research work is to develop the missing knowledge to fill in the highlighted gap in the production of Al-alloy foams of uniform closed-cell

Ashby, M.F., Evans, A., Fleck, N.A., Gibson, L.J., Hutchinson, J.W. & Wadley, H.N.G. (2000).

Babcsán, N., García-Moreno, F. & Banhart, J. (2005). Metal foams—high temperature colloids: Part I. Ex situ analysis of metal foams. 261(1-3):123-130, 2005.

*Metal foams — a design guide,* Butterworth-Heinemann, ISBN 0-7506-7219-6, London,

Fig. 25. (a) Absorbed energy per volume in a certain strain interval is the area under the stress-strain curve. (b) Absorbed energy curves of AlSi7 foam under compressive loading at

)

15

20

25

AlSi7 foams 0.47 g/cm<sup>3</sup> 0.54 g/cm<sup>3</sup> 0.55 g/cm<sup>3</sup> 0.58 g/cm<sup>3</sup> 0.58 g/cm<sup>3</sup> 0.58 g/cm<sup>3</sup> 0.60 g/cm<sup>3</sup>

of the energy absorption with increasing foam density is clearly obvious (Fig.25b).

0 20 40 60 80 100

stressC2MPa

Strain (%)

structures and transfer it to industry.

0

different densities.

**5. Challenges** 

**6. References** 

England

20

40

Stress (MPa)

60

80

100

Area=474,3 dx=50,00

Area=544,1 dx=55,98

latter is almost totally reversible, and occurs uniformly throughout the sample. After reaching the elastic limit, the collapse of the cells starts, mostly by distortion (stretching), rotation and/or sliding of the edges and cell walls, with permanent deformation (points C in Fig. 24b). A progressive collapse of the cells was observed in the "plateau" stage of the loaddisplacement curve (points C to F, in Fig. 24b). This deformation is not uniform due to the irregular structure of the foam (pore size distribution, thickness of cell walls, etc.) (Fig. 20a). The slope that characterises this region may be related to the compression of fluid trapped inside the cells, or due to tensile stress in the cell walls. The slope increases with increasing density of the foam. The shape of collapsed cells is very different from its original shape, as it contains bended and distorted cell walls that may even touch each other. However, in general, the fracture of cell walls does not occur. The initial collapse begins in a small group of cells in the region with the lowest local density of the sample. Collapse does not occur in all cells (points E, in Fig.24b), starting in the cells that are less resistant or with higher loads. The collapse of a cell induces the collapse of neighbouring cells. Moreover, the collapse of neighbouring cells evolves in successive layers and eventually leads to the formation of a single deformation band (points E in Fig.24b).

Al-alloy foams are often used as filler material in lightweight structures subject to crash and/or high velocity impact or as thermal/acoustic insulation devices. The energy

Fig. 24. Load-displacement curve of a AlSi7 foam under compressive loading. Cellular structures of deformed AlSi7 foam samples at different cross-head displacements.

latter is almost totally reversible, and occurs uniformly throughout the sample. After reaching the elastic limit, the collapse of the cells starts, mostly by distortion (stretching), rotation and/or sliding of the edges and cell walls, with permanent deformation (points C in Fig. 24b). A progressive collapse of the cells was observed in the "plateau" stage of the loaddisplacement curve (points C to F, in Fig. 24b). This deformation is not uniform due to the irregular structure of the foam (pore size distribution, thickness of cell walls, etc.) (Fig. 20a). The slope that characterises this region may be related to the compression of fluid trapped inside the cells, or due to tensile stress in the cell walls. The slope increases with increasing density of the foam. The shape of collapsed cells is very different from its original shape, as it contains bended and distorted cell walls that may even touch each other. However, in general, the fracture of cell walls does not occur. The initial collapse begins in a small group of cells in the region with the lowest local density of the sample. Collapse does not occur in all cells (points E, in Fig.24b), starting in the cells that are less resistant or with higher loads. The collapse of a cell induces the collapse of neighbouring cells. Moreover, the collapse of neighbouring cells evolves in successive layers and eventually leads to the formation of a

Al-alloy foams are often used as filler material in lightweight structures subject to crash and/or high velocity impact or as thermal/acoustic insulation devices. The energy

Cross-head velocity: 1.0 mm/min

**E**

Temperature: 22ºC Samples number: 4 AlSi7 Foam

I II

(a)

B: 1mm C: 2mm D: 4 mm E: 8 mm F: 16mm

0 2 4 6 8 10 12 14 16 18 20 22 24 26

**F**

III

Displacement [mm]

Fig. 24. Load-displacement curve of a AlSi7 foam under compressive loading. Cellular structures of deformed AlSi7 foam samples at different cross-head displacements.

(b)

single deformation band (points E in Fig.24b).

0

**A B** **C D**

5

10

15

Load [KN]

20

25

30

absorption capability of these foams can be well estimated from the stress-strain compression behaviour of the material which is estimated from the area under the stressstrain curve (Fig. 25a). As foam materials exhibit a constant stress "plateau" they can absorb higher levels of energy than dense aluminium alloys. Most of the absorbed energy is irreversibly converted into a plastic deformation energy which is a further advantage of foamed Aluminium. For the same stress level, the dense material is deformed in the regime of reversible linear-elastic stresses, releasing most of the stored energy when the load is removed. Al-alloy foams exhibit higher energy absorption capabilities (Fig. 25). The increase of the energy absorption with increasing foam density is clearly obvious (Fig.25b).

Fig. 25. (a) Absorbed energy per volume in a certain strain interval is the area under the stress-strain curve. (b) Absorbed energy curves of AlSi7 foam under compressive loading at different densities.

#### **5. Challenges**

Despite its technological advances, the metallic foam formation is not problem free and still poses challenges. Questions related to a very hot topic (i.e. the control of the pores size and shape of metal foams) that is seen with alacrity by the scientific community due to potential applications of these materials in the transport industry, are highlighted and discussed in detail. The key-question is how to produce metal foams, in series, achieving uniform cellular structure, in order to improve the manufacture reproducibility and to control foam architecture. A key goal of this group research work is to develop the missing knowledge to fill in the highlighted gap in the production of Al-alloy foams of uniform closed-cell structures and transfer it to industry.

#### **6. References**


Aluminium Alloy Foams: Production and Properties 71

Garcia-Moreno, F. , Babcsan, N. & Banhart, J. (2005). X-ray radioscopy of liquid metal foams:

Haesche, M., Lehmhus,D., Weise, J., Wichmann, M. & Mocellin, I. C.M. (2010). Carbonates

Helwig, H.-M., Garcia-Moreno, F. & Banhart, J. (2011). A study of Mg and Cu additions on

Ibrahim, A., Körner, C. & Singer, R.F. (2008). The effect of TiH2 particle size on the

Lehmhus, D. & Busse, M. (2004). Potential New Matrix Alloys for Production of PM

Jiménez, C., García-Moreno, F., Mukherjee, M., Görke, O., Banhart, J. (2009). Improvement

Markaki, A.E. & Clyne, T.W. (2001). The effect of cell wall microstructure on the

Matijasevic, B. & Banhart, J. (2006). Improvement of aluminium foam technology by

Matijašević, B., Banhart, J., Fiechter, S., Görke, O., Wanderka, N. (2006), Modification of

Miyoshi, T., Itoh, M., A. kiyama, S. & Kitahara, A. (2000). Alporas aluminum foam:

Mukherjee, M (2009). *Evolution of metal foams during solidification*, *Technischen Universität* 

Proa-Flores, P.M. & Drew, RAL, (2008).Production of Aluminum Foams with Ni-coated

Kennedy, A. R. (2002). Effect of compaction density on foamability of Al-TiH2 powder compacts. *Powder Metallurgy*, Vol. 45, No. (1), pp. 75-79, ISSN 0032-5899 Kennedy, A. R. (2004). Effect of foaming configuration on expansion. *Journal of Materials* 

*Science & Technology,* Vol. 26, No. (9), pp. 845-850, ISSN 1005-0302

pp. 247–256, ISSN 0334-6455

pp. 5227-5236, ISSN 0022-2461.

pp. 1677-1686, ISSN 1359-6454

1656.

1359-6462

*Berlin* 

1438-1656

Vol. 10, pp. 845–848, ISSN 1438-1656

Vol.61, No.5, pp., 552–555, ISSN 1359-6462

54, N.o 7,pp. 1887–1900, ISSN 1359-6454

Vol. 2, No. 4, pp. 179-183, ISSN 1438-1656

*Science.* Vol. 39, pp. 1143-1145, ISSN 0022-2461

influence of heating profile, atmosphere and pressure. *Colloids and Surfaces A: Physicochemical and Engineering Aspects*, Vol. 263, pp. 290–294, ISSN 0927-7757 Gibson, L.J. & Ashby, M.F. (1997). *Cellular solids – Structure and properties*, Second Edition, Cambridge University Press, ISBN 0-521-49911-9, Cambridge, United Kindgom Gokhale, A., Sahu, S.N., Kulkarni, V.K.W.R., Sudhakar, B., Rao & N. Ramachandra (2007)

Effect of Titanium Hydride Powder Characteristics and Aluminium Alloy Composition on Foaming. *High Temperature Materials and Processes*. Vol. 26, No.4,

as Foaming Agent in Chip-based Aluminium Foam Precursor. *Journal of Materials* 

the foaming behaviour of Al–Si alloys. *Journal of Materials Science*, Vol. 46, No. (15),

morphology of Al-foam produced by PM process. *Advanced Engineering Materials*,

Aluminium Foams. *Advanced Engineering Materials*, Vol. 6, pp. 391-396, ISSN 1438-

of aluminium foaming by powder consolidation under vacuum. *Scripta Materialia*,

deformation and fracture of aluminum-based foams, *Acta Materialia,* Vol. 49, N.o 9,

tailoring of blowing agent. *Scripta Materialia*, Vol. 54, No. (4), pp. 503-508, ISSN

titanium hydride for improved aluminium foam manufacture, *Acta Materialia*, Vol.

production process, properties, and applications. *Advanced Engineering Materials*,

TiH2 Powder. *Advanced Engineering Materials*, Vol. 10, No. (9), pp. 830-834, ISSN


Babcsán, N., García-Moreno, F. & Banhart, J. (2007). Metal foams - High temperature

Banhart, J. (2001). Manufacture, characterisation and application ofcellular metals and metal foams *Progress in Materials Science* Vol. 46, No. (6), pp. 559–632, ISSN 0079-6425 Banhart, J. (2006), Metal foams: production and stability, *Advanced Engineering Materials*, Vol.

Banhart, J. (2008). Gold and gold alloy foams. *Gold Bulletin,* Vol. 41, No. (3), pp. 251-256,

Banhart, J. & Baumeister, J. (1998). Deformation characteristics of metal foams. *Journal of* 

Baumgärtner, F., Duarte, I. & Banhart, J. (2000). Industrialisation of P/M foaming process. *Advanced Engineering Materials*, Vol.2, No.4, pp. 168-174, ISSN 1438-1656 Belkessam, O. & Fritsching U. (2003). Modelling and simulation of continuous metal *foaming* 

Brunke, O. & Odenbach, S. (2006). In situ observation and numerical calculations of the

Campana, F. & Pilone, D. (2008). Effect of wall microstructure and morphometric

Cambronero, L.E.G., Ruiz-Roman, J.M., Corpas, F.A. & Ruiz Prieto, J.M. (2009).

Duarte, I. & Banhart, J. (2000). A study of aluminium foam formation – kinetics and microstructure. *Acta Materialia*, Vol.48, No.9, pp. 2349-2362, ISSN 1359-6454 Duarte, I., Banhart, J. Ferreira & M. Santos. (2006) Foaming around fastening elements.

Duarte, I., Santos, M. & Vide, M. (2008). Processo contínuo de produção de peças e

Duarte, I. (2005), Espumas Metálicas: Processos de fabrico, Caracterização e Simulação numérica. PhD thesis. FEUP- Faculty of Engineering of University of Porto, Porto. Duarte, I., Teixeira-Dias, F., Graça, A. & Ferreira, AJM. (2010). Failure Modes and Influence

Fang, J., Yang., Z., Zhang, H. &Ding, B. (2005). The coating process of silica film on TiH2

protótipos em espumas metálicas. Ingenium, II série, Março/Abril 2008, Vol. 104,

of the Quasi-static Deformation Rate on the Mechanical Behavior of Sandwich Panels with Aluminum Foam Cores. *Mechanics of Advanced Materials and Structures*,

particles and gas release characteristic. *Chemical Engineering Science*, Vol. 60, pp.

*Materials Science Forum*, Vol. 514-516, pp. 712-717, ISSN 0255-5476

*process. Modelling and Simulation in Materials Science and Engineering*, Vol. 11, No.6,

evolution of metallic foams. *Journal of Physics: Condensed Matter*, Vol. 18, pp. 6493–

parameters on the crush behaviour of Al-alloy foams. *Materials Science &* 

Manufacturing of Al–Mg–Si alloy foam using calcium carbonate as foaming agent. *Journal of Materials Processing Technology*, Vol. 209, pp. 1803–1809, ISSN 0924-0136 Degischer, H.P. & Kriszt, B. (Ed(s).) (2002). *Handbook of Cellular Metals*, Wiley-VCH, ISBN 3-

*Materials Science*, Vol. 33, pp. 1431–1440, ISSN: 0022-2461

*Engineering A*, Vol. 479, No. (1-2), pp. 58-64, ISSN 0921-5093

7757

ISSN 0017-1557

8, No. (9), pp. 781–794, ISSN 1438-1656

pp. 823–837, ISSN 0965-03932003.

6506, ISSN 0953-8984

527-30339-1, Weinheim

pp. 78-80, ISSN 0870-5968I.

845-850, ISSN 0009-2509

Vol.17, No. (5), pp. 35-342, ISSN 1537-6494

colloids. Part II: In-situ analysis of metal foams. *Colloids and Surfaces A: Physicochemical and Engineering Aspects*, Vol. 309, No. (1-3), pp. 254–263, ISSN 0927-


**4** 

Burcu Ertuğ *Gedik University,* 

*Turkey* 

**The Fabrication of Porous Barium Titanate Ceramics via Pore-Forming Agents (PFAs)** 

*Department Of Metallurgical & Materials Engineering, Yakack/Kartal, Istanbul* 

The perovskite family includes many titanates used in various electroceramic applications, for example, electronic, electro-optical, and electromechanical applications of ceramics. Barium titanate, perovskite structure, is a common ferroelectric material with a high dielectric constant, widely utilized to manufacture electronic components such as mutilayer capacitors (MLCs), PTC thermistors, piezoelectric transducers, and a variety of electro-optic

For positive temperature coefficient of resistance (PTCR) and humidity/gas sensing applications of BaTiO3, a porous microstructure is required. When oxygen is adsorbed on the grain boundaries, PTCR effect is enhanced. Thus a porous structure is needed for PTCR properties. Besides, porous ceramics can also be used for humidity/gas sensing. The water

A number of routes are employed in order to fabricate porous ceramics for PTCR and/or humidity/gas sensors. Low pressure forming can be used for the production of BaTiO3 ceramics. Forming can be done by unaxial pressing since hot pressing or hot isostatic pressing can eliminate pores. Also several other pressureless forming techniques can be used to make porous bodies. Another method for fabricating porous ceramics is using pore forming agents (PFAs) prior to sintering. PFAs form porosity through the ceramic body by different mechanisms. Different porosifiers that can be used for the production of barium titanate ceramics are starch based, carbon based, metallic or polymeric porosifiers. In the following paragraphs, crystal structure of, donor doping of and the electrical properties of barium titanate is described. Also the concept of porosity is briefly mentioned. Porous ceramics production, pore formers in general and pore formers in BaTiO3 is also explained.

vapour is adsorbed on pores improving the electrical conductivity of the surface.

Finally PTCR and humidity/gas sensing properties of BaTiO3 is described in detail.

ABO3 oxides which adopt the perovskite structure form an important group of compounds possessing many useful and interesting physical properties. The idealized structure is cubic,

**2. Perovskites and barium titanate (BaTiO3)** 

**1. Introduction** 

devices (Wang, 2002).

**for Thermistor and Sensor Applications** 

