Compressive behaviour of open-cell Al–Si alloy foam produced by infiltration casting

Abstract

Open-cell Al–Si alloy foams with different densities were manufactured by vacuum infiltration. The foam with smaller NaCl as space holders showed a mixed scaffold structure consisting of abundant struts and cell walls. It was also found that the open-cell Al–Si alloy foam showed various defects containing micropores, microcracks and missing cell walls. The compression results indicated that the open-cell Al–Si alloy foams exhibited smoother stress–strain behaviour and higher compressive strength than that of previous studies, despite a larger number of defects in brittle Al–Si matrix. The successive occurrences of multiple deformation bands owing to the mixed structure and various defects were responsible for the smooth compression stress–strain behaviour.

Keywords

Porous materials space holder compressive properties deformation mechanisms

Introduction

Porous metallic foams have a special open-cell structure with strong connectivity, which endow with multifunctional and integrated performance, such as their light weight, energy absorption, sound absorption, heat dissipation, and electromagnetic shielding [1–5]. Metallic foams are applied in the aerospace, automotive industry, defence engineering, building materials, and sound absorption materials [6–15]. The feasible methods for manufacturing porous metallic foams are powder metallurgy, electro-deposition, investment casting, space holder filling, infiltration casting, etc. [16–20]. Among all of these methods currently employed, powder metallurgy and infiltration casting have been more widely used to manufacture porous aluminum foam due to their experimental simplicity and controllability of pore distribution, morphology, and size. For the powder metallurgy method, with the small addition of sintering aids and binders, aluminum powders are mixed with space holders in different proportions to be compacted to samples. Melt or sinter aluminum powder and remove space holders after the sample is cooled to obtain open-cell aluminum foam. The frequently employed space holders in powder metallurgy are NaCl [13,21,22], CaCl₂ [23], carbamide [24–26], and sucrose [24,27,28]. However, the powder metallurgy method commonly uses pure aluminum powder as raw material. The process parameters during manufacturing powder vary with the different powders, and the diversity of material properties is greatly limited. Besides, it's difficult for the powder metallurgy method to achieve industrial samples with large sizes due to the limitation of experimental conditions.

The infiltration casting as one of the earliest and most applied methods for open-cell aluminum preparation, uses similar space holders such as NaCl, CaCl₂, MgSO₄, carbamide, and sucrose. The space holders are filled in the mould, and the melted aluminum liquid is infiltrated into the interstices of space holders under different pressures. Depending on the various types of pressure, infiltration casting can be divided into vacuum infiltration method, centrifugal infiltration method, and gravity infiltration method. The infiltration casting process provides a high degree of control over the topology of foam. Chang et al. [29] reported the open-cell aluminum foams with a pore size of 4.5, 3.5, 2.0, and 0.5 mm manufactured by gravity infiltration. The increasing gravity coefficient from 300 to 900 significantly enhanced the structural integrity of pores and the thickness of the cell struts of aluminum foams. Wan et al. [20] reported open-cell aluminum foams with high porosity via the high-temperature deformation of space holders CaCl₂. Besides, several reinforcing materials such as hollow spheres [30–33], ceramic spheres [34–36], polymer [37], and nanomaterials [38,39] have been added to the metal matrix applying powder metallurgy and infiltration casting method, forming metal matrix syntactic foams (MMSFs). The MMSFs feature a lightweight structure that combines high strength with the excellent cushioning and energy absorption properties of metallic foam, which results in promising research and application possibilities.

Based on the above literature, the number of publications on controlling the topology of foam is limited. Therefore, the main objective of the proposed work is to study the influence of space holders with different sizes and morphologies on the microstructure and deformation mechanism of open-cell Al–Si foams. In this work, the NaCl and CaCl₂ particles with different morphologies and sizes are selected as space holders to design the different microstructure of open-cell Al–Si foams by vacuum infiltration. This paper further provides a comparative study about the effect of space holders’ size and morphology on the microstructure, compression properties, and deformation mechanism.

Methodology

Materials and experiments

In the current study, the commercial cast Al–Si alloy (ZL102, Fushun Aluminum Co., Ltd., Fushun, China) was employed as the metal matrix material for manufacturing aluminum foam; its chemical composition is listed in Table 1. Figure 1 shows the X-ray diffraction patterns of the cast Al–Si alloy, mainly presenting two diffraction peaks of α-Al and β-Si.

Table 1

Composition of aluminum-silicon alloy used in the present work.

Element	Al	Si	Fe	Mn	Cu	Mg
Contents/wt.%	Bal.	9.75	0.03	0.004	0.0043	0.005

Two different water-soluble space holders were employed, namely, NaCl particles (Shenyang Fengxing Salt Manufacturing Co., Ltd, Shenyang, China) with purity over 98.5% and CaCl₂ particles (Anhui Yihua Packaging Co., Ltd, Fuyang, China) with purity over 96%. NaCl particles were sieved between 0.15–0.25 mm and 0.25–0.6 mm in diameter, while the CaCl₂ particles were sieved about 2 mm in diameter. The morphology of the selected space holders as characterised using a scanning electron microscope (SEM) is depicted in Figure 2. NaCl particles with small sizes (Figure 2(a)) were mainly irregular shape, CaCl₂ particles (Figure 2(c)) were regular spheres, and NaCl particles with large sizes (Figure 2(b,d)) owned both shapes. ANS foam and ANL foam represented open-cell aluminum foam prepared with NaCl as space holders with small and large pores, respectively, and AC foam represented the sample prepared using CaCl₂ as space holders.

The vacuum infiltration method consists of majorly four steps including filling of space holders, mould insulation, vacuum infiltration casting of liquid aluminum and cooling. A schematic of the open-cell aluminum foam vacuum infiltration manufacturing process is shown in Figure 3. Complete process of the vacuum infiltration method is elaborated as following. Initially, poured about 300 g of space holders into the mould and stacked them into cylinders with 10 cm in diameter and 5 cm in height. Then the mould filled with particles was placed into the resistance furnace to keep the temperature at about 500°C for an hour. At the same time, the aluminum ingots of about 350 g were heated to 670°C and kept at this temperature for more than half an hour. After the end of heat preservation, cast the molten aluminum on the space holders to form a sealing layer and opened the vacuum equipment. The vacuum level was varied between 0.4 and 0.9 bar, which depended on the size and morphology of the particles. Castings were solidified in the mould, without additional measures to control the direction or speed of solidification. Finally, rinsed the sample with flowing deionised water, dissolved the particles in the pores, and obtained the open-cell aluminum foam. The morphologies of the cut samples (10 × 10 × 10 mm) of ANS, ANL, and AC foams are depicted in Figure 4. The macroscopic morphology, pore structure, and microstructure were observed using the stereomicroscope and scanning electron microscopy (SEM).

Quasi-static compression

The ANL, ANS, and AC foam samples used for the compression test were ground into 15 × 15 × 15 mm, and the typical characteristics of compression samples are shown in Table 2. The relative densities ρ_R were calculated from their densities of samples and the matrix material using the following equation:

ρ_{R} = \frac{ρ_{S}}{ρ_{M}}

(1)

where ρ_S and ρ_M were the densities of the sample and the matrix material respectively. The density distribution is not uniform even within the same sample, resulting from the heterogeneous and discontinuous porous structure inside the metallic foams. Thus, four compressed samples were obtained by random cutting within each of the different open-cell aluminum foams to reduce experimental uncertainty and avoid contingency. The densities of compressed samples of the ANL, ANS, and AC foams ranged from 0.63 to 0.79 g/cm³, 0.82 to 0.94 g/cm³, and 1.01 to 1.10 g/cm³, respectively.

Table 2

Densities of ANS, ANL, and AC foam samples.

Sample	Material	Density (g/cm³)	Relative density
ANL-0.23	ANL foam	0.63	0.23
ANL-0.26	ANL foam	0.70	0.26
ANL-0.27	ANL foam	0.73	0.27
ANL-0.29	ANL foam	0.79	0.29
ANS-0.30	ANS foam	0.82	0.30
ANS-0.30	ANS foam	0.82	0.30
ANS-0.33	ANS foam	0.88	0.33
ANS-0.35	ANS foam	0.94	0.35
AC-0.37	AC foam	1.01	0.37
AC-0.40	AC foam	1.07	0.40
AC-0.41	AC foam	1.10	0.41
AC-0.41	AC foam	1.10	0.41

Quasi-static compression tests were carried out at a constant displacement rate of 0.025 mm/s by CMT5105 universal mechanical testing machine with a maximum test force of 50 kN. The relative error of the test force and displacement value is less than 0.5% while the control accuracy is less than 0.1%. The compressive stress–strain curve was calculated according to the original cross-section and height of each sample. The yield strength, plateau stress, and densification strain were to refer to ISO 13314 standard [10,13] and all data processing and analysis were based on the following method unless otherwise specified. The yield strength σ_y was defined as the stress at the strain of 0.05. The plateau stress σ_p was the arithmetical mean of the stress corresponding to 0.2–0.4 strain. The densification strain ε_d was the strain corresponding to 1.3 times plateau stress. The energy absorption W was the integral under the compressive stress–strain curve;

W = \int σ d ε

(2)

where σ and ε were compressive stress and strain. The energy absorption efficiency I was calculated by the following formula:

I = \frac{\int_{0}^{ε} σ d ε}{σ ε}

(3)

The hardening rate d

σ

/dε can be obtained by the derivative of the stress–strain curve.

Furthermore, in this work, a relatively reasonable alternative method was developed to characterise the compressive properties of porous metallic foams, especially in the smooth stress–strain response. Figure 5 shows compressive performance indexes for open-cell aluminum foam. Yield strength, compressive strength, densification strain, energy absorption, energy absorption efficiency, and hardening rate were determined by numerical methods combined with actual physical meaning on the basis of previous work [6–11]. As Figure 5 shown, the yield strength σ_y was defined as the stress at strain 0.05. The compressive strength σ_c was the stress corresponding to the point where the second derivative of the stress–strain curve was 0, i.e. the inflection point of the curve. The densification strain ε_d was the strain corresponding to the intersection of the tangent of the point and the x-axis when the slope of the stress–strain curve coincided again with the slope of the elastic stage.

Results

Structural characteristics

Figure 6(a–f) displays the series of pore structures of AC, ANL and ANS foams. Figure 6(a,d) exhibits the pore structure of AC foams. The pores were nearly spherical with thicker cell walls, evenly distributed in the matrix, presenting typical open-cell wall structure. Some connectivity was found through lots of micropores growing on the cell walls. Figure 6(b) shows the pore structure characteristics of the ANL foam sample. The pores were mostly uneven in shape, with a few spherical pores, which was consistent with the morphology of space holders. A special scaffold structure was observed, comprised mostly of struts and a few cell walls. Plenty of defects, such as missing cell walls, microcracks, and micropores, destroyed the integrity of the cell wall and formed this mixed scaffold structure with stronger connectivity. The different wettability between molten liquid aluminum and solid space holders during manufacturing, which caused insufficient infiltration, was responsible for the formation of defects and thin cell walls. Figure 6(e) depicts the typical struts of ANL foam. The pore structure composition of ANS foam observed in Figure 6(c) was comparable to ANL foam, but with smaller pores. The missing cell walls are magnified in Figure 6(f). Despite having more struts and stronger connectivity, the pore structure of ANS and ANL foams was still different from the network structure [29], which also guaranteed higher strength.

Figure 7 presents the microstructure of cell walls in ANL and ANS foams. The primary aluminum phase and Al–Si eutectic constituted the predominant microstructure on cell walls. The obvious corrosion gaps resulted from the corrosion of chlorine ions in the process of removing space holders could be observed on the surface of cell walls and struts, which gave rise to the separate eutectic Si lamella. EDS measurements were carried out for region A (Figure 7(a)) and region B (Figure 7(d)). As shown in Figure 8(a), the dendritic slender eutectic silicon alternately grew on α-Al, with each layer having a width of about 2 μm. The line-scan result shown in Figure 8(b) further verified that Al–Si eutectic was the primary microstructure of cell walls. Besides, a few of white flake and acicular compounds embedded in the aluminum matrix are observed in Figure 9(a,b), with a length of about 10 μm. The EDS results signified that the compounds contained Fe, and the element composition is listed in Table 3.

Table 3

Composition of different areas in Figure 9.

	Mass fraction (wt.%)
Area	Al	Si	Fe	O
C	53.79	25.50	20.71	0
D	49.88	33.41	16.72	0
E	71.07	14.73	9.91	4.29
F	61.44	18.61	14.68	5.27

Mechanical performance

Compression properties

Figure 10 shows the stress–strain curves of the ANL, ANS, and AC foams for different relative densities. And the smooth compression curves presented significant characteristics without peak stress, appearing steady increasing hardening behaviour during the plateau stage. The stress–strain curves could be divided into an initial elastic stage, a plastic plateau stage, and a densification stage. At the beginning of compression, the stress increased quickly with the strain. However, the elastic stage ended (yielded) prematurely due to the special pore structure resulted from defects. And the occurrence of plastic deformation caused the slope of the first stage curve to decrease continuously. Furthermore, followed by the plateau stage, the transition of the compression curve forms the initial stage to the plateau stage was smooth with no stress drop. During the plateau stage, the stress–strain curve didn't maintain a fluctuant stress platform [40,41], but kept rising at a small growth rate, illustrating the hardening behaviour of obvious and steady growth. With continuous compression, the plastic modulus of the materials gradually increased. At the end of the plateau stage, the holes of the foams appeared to be compacted, and extrusion and friction occurred between the cell walls, resulting in a sharp rise in stress, which was the densification stage of deformation.

Figure 10 also signifies that the greater the relative density of the sample, the greater its yield strength and plateau stress. AC foams exhibited greater yield strength and plateau stress. And the stress–strain curves of ANS and ANL foams ended the elastic stage and entered the plateau stage earlier.

Figure 11 statistically present the average yield strengths, average plateau stress, and average densification strains of ANL, ANS, and AC foams. The densification strains of all compressed samples fluctuated in a close range of 0.40–0.44. The AC foams exhibited the greatest performance with an average yield strength of 20.49 MPa and average plateau stress of 33.64 MPa. As the average relative density of open-cell aluminum foam increased from 0.26 to 0.40, the average yield strength increased from 7.34 to 20.49 MPa, while the average plateau stress grew from 14.41 to 33.64 MPa. It could be concluded that the yield strength and plateau stress of the compressed specimen increased with the increase of relative density. The sample with larger relative density had higher aluminum content, thicker cell walls and struts, and a stronger ability to withstand deformation. Numerous studies had shown that relative density was one of the most influential factors in the compressive properties of open-cell foams. Based on the research in Refs. [40–42], the strain for the onset of densification, for a porous metal with a density of 0.83 g/cm², was estimated to be in the range of 0.50–0.56. This prediction was broadly in keeping with the densification stage of 0.49–0.52 calculated by the method proposed in this work for characterising the compressive performance, which also verified the feasibility of this data processing method.

The fracture morphology of ANL foams is displayed in Figure 12. The smooth cleavage fractures appeared under the shear stress, and there were few tear marks between the cleavage planes, appearing obvious brittle fracture characteristics. The open-cell foams with brittle fracture characteristics showed smooth stress–strain curves in Figure 10, a similar phenomenon could be observed in Ref. [43].

Energy absorption

Figure 13 depicts the energy absorption W and energy absorption efficiency I of ANL, ANS, and AC foams. The energy absorption increased with increasing stress. The energy absorption calculated by Equation (2) depended on the stress–strain curve, thus, the AC foams had higher energy absorption compared to ANL and ANS foams. Similarly, samples fabricated by the same space holders with higher relative density had higher energy absorption attributed to their higher yield stress and plateau stress.

The energy absorption efficiency obtained by Equation (3) appeared two distinct stages, the ascent stage, and the descent stage. Due to the continuously increasing hardening behaviour during the plateau stage, the energy absorption efficiency curve of ANL, ANS, and AC foams cannot maintain a stable and high-efficiency platform at the peak after experiencing the rising stage, which was also related to the special deformation modes brought by the defects. AC-0.37 foam reached the highest energy absorption efficiency of 0.82 when the strain was about 0.2. Compared to AC foams, though ANL and ANS foams had lower peak energy absorption efficiency, they all reached the peak energy absorption efficiency sooner. ANL-0.23 foam, the sample with minimum relative density, the earliest (ε = 0.134) reached the highest energy absorption efficiency of 0.747. In the initial stage of deformation, the energy absorption efficiency of ANL and ANS foams was higher than that of the AC foams, which was due to the earlier end of the elastic stage.

Discussion

Compressive behaviour sensitivity

The relationship between the plateau stress and the density of open-cell foam was given by the Gibson-Ashby model [6] as

\frac{σ}{σ_{y s}} = C (\frac{ρ_{S}}{ρ_{M}})^{n}

(4)

where σ was the plastic collapse stress of metallic foams, σ_ys was the yield strength of the matrix materials, n was equal to 1.5, where, as always, C was constant. Equation (4) could be simplified into the following Equation (5) after the logarithm.

log σ = n \log ρ_{S} + C^{'}

(5)

where n was 1.5 in the Gibson-Ashby model for open-cell foam, C′ was the constant. And in this study, σ was defined as plateau stress σ_p.

The relation between plateau stress and the densities of ANL, ANS, and AC foam samples in this work is characterised in Figure 14. The slope of the curve, n in Equation (5), represented the sensitivity to the compression behaviour of the material. The correlation coefficient R² of the line fitted by the experimental data was 0.97, indicating the suitability of the Gibson-Ashby model for ANL, ANS, and AC foams. A reasonable explanation was that the pore structure in this work, especially in ANS and ANL foam, was more in the forms of struts and edges on account of many defects, e.g. missing cell walls and microcracks, which was highly consistent with the Gibson-Ashby model of prismatic model. However, the slope of the fitted line was quite different from that of the Gibson-Ashby model, implying that the compression samples of ANL, ANS, and AC foams were more sensitive to compressive behaviour. On the one hand, there was a deviation during the definition of parameters. In the Gibson-Ashby model, the yield stress of open-cell foam was specified as the fully plastic collapse of cell structure [6]. While, in this work, the stress–strain curve of the samples did not appear evident limit of elastic stage and peak stress, and the plateau stress σ_p was used to represent the plastic collapse stress σ in Equation (5). On the other hand, the actual pore structure of the samples of ANL, ANS and AC foams was more complex than that of Gibson-Ashby model. In addition to struts, the pore structure of the samples contained a few cell walls, and these diverse forms constituted the mixed scaffold structure, which enhanced the compressive strength under the same density compared to the Gibson-Ashby model.

Figure 15 depicts the results of some previous studies on open-cell Al foams and Zn foams [11,24,40,44]. The relationship between the stress corresponding to the complete plastic collapse of open-cell foams and the foam densities with different materials was established according to the Gibson-Ashby model. The details of the sample preparation are arranged in Table 4, and the slope n of each line fitted according to the experimental compression data was also counted. As could be observed from Figure 15, two different areas of data accumulation were marked with the data in the circle on the left coming from the compression data of open-cell Al foams and the open-cell Zn foams on the right. Except for the findings from Ref. [40], the sensitivity of both Al and Zn foams to compression behaviour was higher than the Gibson-Ashby model. It could also be analysed that, for the same density, the closer the data was to the left upper side of the figure, the greater strength of the material was. And obviously, in contrast to open-cell Zn foams, the Al foams exhibited better compression properties at the same density. The ANL, ANS and AC foams prepared by vacuum infiltration in this work signified the best compression performance among all experimental data. The fitting degree of the Gibson-Ashby model to the actual pore structure of the materials had been discussed, and the correlation coefficient R² of the fitted line was used as the standard for the discussion. In Ref. [11], the fitted line from #4 had a higher R²= 1.00 than the fitted line from #5 (R²= 0.99), indicating that Zn foams prepared with NaCl of 2–2.5 mm as space holder had the pore structure closer to Gibson-Ashby model compared with Zn foams prepared with NaCl of 4–4.5 mm as a space holder. The pore structure of the foam material will appear more struts rather than a complete cell wall using the smaller space holders, which was also confirmed in this study.

Table 4

Structural parameters of various open-cell metal foam samples.

Material	Preparation method	Particle	Pore size (mm)	Density (g/cm²)	Relative density	n in Equation (4)	References
Al	Powder metallurgy	Carbamide	1.4	0.4–1.4	0.2–0.5	2.4	[24]
Al	Powder metallurgy	Sucrose	1.3	0.5–1.3	0.2–0.5	2.7	[24]
Al alloy	Vacuum infiltration	NaCl	2.0–3.0	0.7–0.9	0.3	1.3	[40]
Zn	Vacuum infiltration	NaCl	2.0–2.5	2.4–2.9	0.4	3.0	[11]
Zn	Vacuum infiltration	NaCl	4.0–4.5	2.1–2.5	0.3	2.2	[11]
Zn alloy	Centrifugal infiltration	NaCl	0.4–0.9	1.7–2.2	0.3–0.4	2.4	[44]
Al alloy	Vacuum infiltration	NaCl/CaCl₂	0.1–0.6	0.6–1.1	0.2–0.4	2.4	Present work

Smooth stress–strain response

As illustrated in Figure 10, the smooth compressive stress–strain curve was the result of global deformation or successive occurrences of multiple deformation bands of the structure [45]. The local stress concentration and stress release will cause stress fluctuations. Commonly, a sudden collapse of individual deformation bands will lead to a stress drop on the stress–strain curve. Followed by the yield and collapse of another individual deformation band away from the last one. Such a process appears alternately, causing a significant fluctuation in the compression curve. In this work, the large number of defects inside the open-cell aluminum foam was an essential trigger for the non-synchronous occurrence of multiple deformation bands. The defects took the forms of mixed scaffold structure containing missing cell walls, microcracks, and micropores, independent struts of different sizes, various corrosion present on the cell walls, etc. And these defects developed multiple discrete weaker areas that were more prone to yielding and the formation of deformation bands. During compression, the discrete weaker areas collapsed asynchronously, resulting in multiple deformation bands and maintaining a smooth compression response.

The mechanical response in this work was quite different from that of other studies listed in Table 3 [11,24,40,44]. Sathaiah et al. [24] reported a significant influence of the shape of the space holders for the mechanical properties of the foams in the powder metallurgy method. Compared to Al foams prepared by sucrose, Al foams prepared by carbamide had higher plateau stress (∼3 times), energy absorption (∼3 times), smoother stress–strain curve, and improved young's modulus. Cheneler and Kennedy [40] reported the aluminum syntactic foams and porous aluminum prepared by vacuum infiltration. Though the syntactic foams had a higher yield strength and plateau stress by adding expanded glass particles, the stress–strain curves of the porous foam were smoother and the structure was more efficient. Open-cell Zn-22Al-2Cu foams with pore sizes of 0.42, 0.65, and 0.85 mm fabricated by centrifugal infiltration process were reported by Sánchez-Martínez et al. [44]. Foams with smaller pore sizes had smaller pore struts, improved mechanical properties, and smoother stress–strain curves. The research reported that the stress–strain curves of metal foams mostly exhibited peak yield strength and curve fluctuations. It was further illustrated that the pore structure of ANL, ANS, and AC foams manufactured in this work had played a great role in their mechanical properties.

Hardening behaviour

Figure 16 suggests the relation between hardening rate d $σ$ /dε and strain. The hardening curve could be divided into three stages corresponding to the three stages of stress–strain curves: (1) an initial, significant decreasing stage showed the increasing magnitude of stress decreased with strain until (2) the stage of rising steadily in hardening rate, and finally (3) a densification stage, in which the stress increased steeply. The sharp drop in hardening rate indicated the plastic deformation even at the beginning of compression. It may be associated with weak region such as extremely thin struts and defects. As the hardening rate decreased and began to flatten, it means that the stress–strain curve entered the plateau area. At this stage, the hardening rate was still greater than zero because the samples had a gradually increasing hardening behaviour. The hardening rate continued to decline during the plateau stage and eventually reached the lowest point, which was also the inflection point of the stress–strain curve, and the stress–strain curve became the ‘flattest’, which means the beginning of the full-scale yield of global samples. According to Equation (3), the energy absorption efficiency I reached the maximum shown in Figure 12 meanwhile. After the compression process passed through the inflection point of the stress–strain curve, the main plastic behaviour of the porous structure was gradually transformed from buckling, plastic hinge to fracture, extrusion friction, and finally through continuous hardening into the densification stage.

Deformation mechanism

Figure 17 suggests the typical compression process of ANL foams from the strain of 0.05–0.5. Its special pore structure of mixed scaffold structure was discussed in section ‘structural characteristics’, which caused different deformation mechanisms. Figure 17(a) shows the deformation of the sample at the strain of 0.05. At the beginning of compression, the initial elastic stage, in addition to the elastic deformation of the material, special deformation modes appeared in the ANL foam structure. The enlarged views of areas of I and II are depicted in Figure 18(a). Cell A underwent buckling, which was more common in the plastic plateau stage of classical foam compression [46]. Meanwhile, more struts structure existing in the ANL foam structure occurred with plastic hinges and even the fracture of the cell walls. The appearance of these typical plastic behaviour was due to stress concentration resulted from its extremely thin hole walls and early failure of initial defects, such as microcracks, and missing cell walls. Weaker regions, e.g. thinner hole walls and pore structures with defects, will occur obviousplastic deformation even under little stress, leading to the termination of the elastic stage of ANL foam at the initial stage of deformation. It could explain the phenomenon of no obvious elastic limit of stress–strain curves in Figure 10. In contrast to ANL foams, AC foams grew thicker pore walls, and had a more complete cell wall structure. As shown in Figure 18(a), under the same strain, the global elastic deformation was the main deformation mode of AC foams, accompanied by the buckling of some cell walls, which was attributed to the stress concentration caused by the micropores. Less plastic deformation, without plastic hinge and fracture, resulting in the slower attenuation of the slope of the stress–strain curves in the first stage of AC foams in Figure 10.

Figure 17(b–e) presents the deformation process from ε = 0.05 to ε = 0.30 of ANL foam. It could be observed that there were no obvious deformation bands during compression. After the failure of the more easily deformed struts structure, due to the rearrangement of stress caused by local deformation, the surrounding struts structure yielded immediately. Such a process randomly took place throughout the foam structure, with plenty of micro deformation bands randomly forming and collapsing asynchronously, leading to global deformation [45]. As Figure 18(b) suggested, AC foam had a different compression behaviour of producing an obvious local deformation band during this stage. However, unlike the deformation bands created by simultaneously collapsed cells, the deformation band of AC foams was gradually shaped due to the continuous plastic deformation of the cell walls with defect concentration. And the difference of deformation behaviour between ANL foams and AC foams led to a smoother stress–strain curve of ANL foam and ANS foam. The size effect in the compression mechanism reflected in this work was similar to the conclusion in Ref. [46]. Apart from buckling and plastic hinge, the more typical deformation mechanism of ANL foam at this stage was fracture of pore walls and further extrusion friction, resulting in growing hardening behaviour of ANL foam in the plateau stage, similar behaviour could be observed in AC foam. Such high strain hardening behaviour of cell struts was also a significant cause of uniform plastic deformation [47]. Figure 18(b) enlarges the extrusion friction behaviour in the region III and IV of ANL foam as well as the fractured cell walls of AC foam.

Figure 17(f) displays the compressive deformation at the strain of 0.5 and Figure 18(c) shows the enlarged view of area V. From this stage, ANL foam began to enter the densification stage, squeezing, sliding, and friction appeared between most hole walls, resulting in rapidly rising stress. The deformation morphology of AC foam at the strain of 0.5 is also depicted in Figure 18(c). The cells within the deformation band were massively broken and some of the cell walls began to squeeze. However, it could be observed that the cells in the deformation band were not completely compacted, while the cells outside the deformation band showed a tendency to be compacted, which signified different evolution of the local deformation band produced by AC foam. With the compression, the cells outside the deformation band also appeared stress concentration owing to defects, e.g. micropores, microcracks, and the stress of the cells in the deformation band was released, resulting in the termination of further stress concentration in the deformation band.

Thus, in spite of the different manifestations of compression behaviour between ANL and AC foams, they both appeared the similar deformation mechanism of global deformation in this study. The asynchronous local stress concentration and relaxation were the crucial explanation for the smooth stress–strain curve without stress drop and stress fluctuation of ANL, ANS, and AC foams.

Conclusions

Open-cell Al–Si alloy foams with different densities ranging from 0.63 to 1.10 g/cm² were successfully manufactured by vacuum infiltration method. NaCl particles with two different sizes (0.15–0.25 mm and 0.25–0.6 mm) as well as sphere CaCl₂ particles (2 mm) were used as space holders. The primary aluminum phase, Al–Si eutectic, and a few of Fe-rich phases could be observed on cell walls. The corrosion of chlorine ions in the process of removing space holders resulted in the corrosion gaps on the surface of cell walls and struts, which gave rise to the separate eutectic Si lamella. Open-cell foams fabricated using NaCl as space holders exhibited mixed scaffold structures consisting of struts and cell walls, and the foam fabricated using CaCl₂ as space holders showed a typical open-cell wall structure. Defects containing micropores, microcracks, missing cell walls, and the corrosion surface of cell walls and struts were found in the open-cell Al–Si alloy foams. Although the cell wall material presented significant brittle fracture, the open-cell Al–Si alloy foams showed smoother stress–strain behaviour and higher compressive strength than that of previous studies. As the average relative density of open-cell aluminum foam increased from 0.26 to 0.40, the average yield strengths increased from 7.34 to 20.49 MPa, while the average plateau stress grew from 14.41 to 33.64 MPa. Moreover, foams manufactured using NaCl and CaCl₂ as space holders both appeared the global deformation owing to mixed structures and various defectes. The cell wall collapsed and struts failed randomly owing to abundant defects in the foam sample, resulting in the formation of multiple random deformation bands. Meanwhile, local stress concentration and relaxation occurred asynchronously, which was responsible for the smooth stress–strain behaviour.

Footnotes

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The datasets generated and/or analysed during the current study are available from the corresponding author on reasonable request.

Authors contributions

Yinzheng Xia: Methodology, Validation, Formal analysis, Investigation, Data curation, Writing - original draft. Jianchao Shi: Investigation, Data curation, Analysed results. Yongliang Mu: Conceptualisation, Methodology, Investigation, Writing - review & editing, Supervision, Project administration, Funding acquisition.

References

Wan

, Liu

, Zhou

, . Fabrication, properties, and applications of open-cell aluminum foams: a review. J Mater Sci Technol. 2021;62:11–24.

Jinnapat

, Kennedy

AR.

The manufacture of spherical salt beads and their use as dissolvable templates for the production of cellular solids via a powder metallurgy route. J Alloys Compd. 2010;499(1):43–47.

Papantoniou

, Markopoulos

, Pantelis

, . Application of aluminium flakes in fabrication of open-cell aluminium foams by space holder method. Materials (Basel). 2018;11(8):1420.

Marx

, Rabiei

Overview of composite metal foams and their properties and performance. Adv Eng Mater. 2017;19:1600776.

Zhou

, Jia

, Zhang

, . Controllable synthesis of Ni/NiO@porous carbon hybrid composites towards remarkable electromagnetic wave absorption and wide absorption bandwidth. J Mater Sci Technol. 2021;87:120–132.

Gibson

, Ashby

MF.

Cellular solids: structure and properties. Cambridge, Cambridge University Press; 1999.

Gama

, Bogetti

, Fink

, . Aluminum foam integral armor: a new dimension in armor design. Compos Struct. 2001;52(3–4):381–395.

Tan

, Harrigan

, Reid

SR.

Inertia effects in uniaxial dynamic compression of a closed cell aluminum alloy foam. Mater Sci Technol. 2002;18(5):480–488.

Lopatnikov

, Gama

, Haque

, . Dynamics of metal foam deformation during Taylor cylinder-Hopkinson bar impact experiment. Compos Struct. 2003;61(1–2):61–71.

10.

, Wang

, Hao

, . Compressive properties of expanded glass and alumina hollow spheres hybrid reinforced aluminum matrix syntactic foams. J Alloys Compd. 2019;821:153233.

11.

Wang

, Zhang

, Zu

, . Compressive properties of biomedical open-cell Zn foam. Mater Res Express. 2019;6:115406.

12.

TJ.

Ultralight porous metals: from fundamentals to applications. Acta Mech Sin. 2002;18(5):457–479.

13.

Surace

, De Filippis

LAC

, Ludovico

, . Influence of processing parameters on aluminium foam produced by space holder technique. Mater Des. 2009;30(6):1878–1885.

14.

Zhao

, Han

, Fung

Optimisation of compaction and liquid-state sintering in sintering and dissolution process for manufacturing Al foams. Mater Sci Eng A. 2004;364(1–2):117–125.

15.

Banhart

, Baumeister

, Weber

Damping properties of aluminum foams. Mater Sci Eng A. 1996;205(1–2):221–228.

16.

, Eifert

, Banhart

, . Metal foaming by a powder metallurgy method: production, properties and applications. Mater Res Innov. 1998;2:181–188.

17.

Maine

EMA

, Ashby

MF.

Applying the investment methodology for materials (IMM) to aluminium foams. Mater Des. 2002;23(3):307–319.

18.

Liu

, Liu

, Shi

, . Adjusting compressive properties of open-cell Mg-Gd-Zn foams by variation of Gd content. Mater Sci Eng A. 2021;820:141562.

19.

Banhart

Manufacture, characterisation and application of cellular metals and metal foams. Prog Mater Sci. 2001;46(6):559–632.

20.

Wan

, Liu

, Zhou

, . Fabrication of high-porosity open-cell aluminum foam via high-temperature deformation of CaCl₂ space-holders. Mater Lett. 2021;284:129018.

21.

Hussain

, Suffin

NSA.

Microstructure and mechanical behaviour of aluminium foam produced by sintering dissolution process using NaCl space holder. J Eng Sci. 2011;7:37–49.

22.

Zhao

, Sun

A novel sintering-dissolution process for manufacturing Al foams. Scr Mater. 2001;44(1):105–110.

23.

Bafti

, Habibolahzadeh

Compressive properties of aluminum foam produced by powder-carbamide spacer route. Mater Des. 2013;52:404–411.

24.

Sathaiah

, Dubey

, Pandey

, . Effect of spherical and cubical space holders on the microstructural characteristics and its consequences on mechanical and thermal properties of open-cell aluminum foam. Mater Chem Phys. 2021;273:125115.

25.

Bafti

, Habibolahzadeh

Production of aluminum foam by spherical carbamide space holder technique-processing parameters. Mater Des. 2010;31(9):4122–4129.

26.

Jiang

, Zhao

, Shi

, . A novel method for making open cell aluminum foams by powder sintering process. Mater Lett. 2005;59(26):3333–3336.

27.

Michailidisa

, Stergioudia

, Tsouknidasb

, . Compressive response of Al-foams produced via a powder sintering process based on a leachable space-holder material. Mater Sci Eng A. 2011;528(3):1662–1667.

28.

Michailidisa

, Stergioudia

Establishment of process parameters for producing Al-foam by dissolution and powder sintering method. Mater Des. 2011;32(3):1559–1564.

29.

Chang

, Gao

, Wang

, . Manufacturing 3-D open-cell aluminum foam via infiltration casting in a super-gravity field. J Mater Process Technol. 2018;252:705–710.

30.

Szlancsik

, Katona

, Bobor

, . Compressive behaviour of aluminium matrix syntactic foams reinforced by iron hollow spheres. Mater Des. 2015;83:230–237.

31.

Kemény

, Leveles

, Bubonyi

, . Effect of particle size and volume ratio of ceramic hollow spheres on the mechanical properties of bimodal composite metal foams. Compos Part A – Appl Sci Manuf. 2021;140:106152.

32.

Szlancsik

, Orbulov

IN.

Compressive properties of metal matrix syntactic foams in uni- and triaxial compression. Mater Sci Eng A. 2021;827:142081.

33.

Katona

, Szlancsik

, Tábicd

, . Compressive characteristics and low frequency damping of aluminium matrix syntactic foams. Mater Sci Eng A. 2019;739:140–148.

34.

Májlinger

, Orbulov

IN.

Characteristic compressive properties of hybrid metal matrix syntactic foams. Mater Sci Eng A. 2014;606:248–256.

35.

Kemény

, Movahedi

, Fiedler

, . The influence of infiltration casting technique on properties of metal syntactic foams and their foam-filled tube structures. Mater Sci Eng A. 2022;852:143706.

36.

Orbulov

, Szlancsik

, Kemény

, . Compressive mechanical properties of low-cost, aluminium matrix syntactic foams. Compos Part A – Appl Sci Manuf. 2020;135:105823.

37.

Marx

, Robbins

, Grady

, . Polymer infused composite metal foam as a potential aircraft leading edge material. Appl Surf Sci. 2020;505:144114.

38.

Yang

, Yang

, Liu

, . High strain rate dynamic compressive properties and deformation behavior of Al matrix composite foams reinforced by in-situ grown carbon nanotubes. Mater Sci Eng A. 2018;729:487–495.

39.

Guo

, Zou

, Shi

, . Compressive properties and energy absorption of aluminum composite foams reinforced by in-situ generated MgAl₂O₄ whiskers. Mater Sci Eng A. 2015;645:1–7.

40.

Cheneler

, Kennedy

AR.

A comparison of the manufacture and mechanical performance of porous aluminum and aluminum syntactic foams made by vacuum-assisted casting. Mater Sci Eng A. 2020;789:139528.

41.

Cheng

, Li

, Chen

, . Compressive properties and energy absorption of aluminum foams with a wide range of relative densities. J Mater Eng Perform. 2018;27:4016–4024.

42.

Ashby

, Evans

, Fleck

, . Metal foams: a design guide. Burlington: Butterworth-Heinemann; 2000.

43.

Demetriou

, Hanan

, Veazey

, . Yielding of metallic glass foam by percolation of an elastic buckling instability. Adv Mater. 2007;19:1957–1962.

44.

Sánchez-Martínez

, Cruz

, González-Nava

, . Main process parameters for manufacturing open-cell Zn-22Al-2Cu foams by the centrifugal infiltration route and mechanical properties. Mater Des. 2016;108:494–500.

45.

Fan

, Zhang

, Gao

, . Deformation mechanisms of spherical cell porous aluminum under quasi-static compression. Scr Mater. 2018;142:32–35.

46.

Taherishargh

, Sulong

, Belova IV , . On the particle size effect in expanded perlite aluminum syntactic foam. Mater Des. 2015;66:294–303.

47.

, Yao

, Liang

, . Deformation mechanisms of closed-cell aluminum foam in compression. Scr Mater. 2010;63(6):629–632.