Processing and Characterization of a New Composite Metal Foam

Materials Transactions, Vol. 47, No. 9 (2006) pp. 2148 to 2153 Special Issue on Porous and Foamed Metals —Fabrication, Characterization, Properties and Applications— #2006 The Japan Institute of Metals

Afsaneh Rabiei, Lakshmi Vendra*1, Nick Reese*2, Noah Young*2 and Brian P. Neville*1

Department of Mechanical and Aerospace Engineering, North Carolina State University, Raleigh, NC 27695-7910, USA

New closed cell composite metal foam has been processed using both casting and powder metallurgy (PM) techniques. The foam is comprised of steel hollow spheres packed into a dense arrangement, with the interstitial spaces between spheres occupied with a solid metal matrix. Using the casting technique, an aluminum alloy inﬁltrates the interstitial spaces between steel spheres. In the PM technique, steel spheres and steel powder are sintered to form a solid, closed cell structure. The measured densities of the Al-Fe composite foam, low carbon steel foam, and stainless steel foam are 2.4, 2.6, and 2.9 g/cm3 with relative densities of 42, 34, and 37%, respectively. The composite metal foams composite materials developed in this study displayed superior compressive strength as compared to any other foam being produced with similar materials. The compressive strength of the cast Al-Fe foam averaged 67 MPa over a region of 10 to 50% strain, while the low carbon steel PM foam averaged 76 MPa over the same strain region, and the stainless steel PM foam averaged 136 MPa over the same region. Densiﬁcation began at approximately 50% for the cast foam and ranged from 50 to 55% for the PM foams. The strength to density ratio of the product of both techniques exceeded twice that of foams processed using other techniques with similar materials. [doi:10.2320/matertrans.47.2148]

(Received February 28, 2006; Accepted April 24, 2006; Published September 15, 2006) Keywords: metal foam, energy absorption, hollow spheres, compression strength, plateau stress

1. Introduction by filling in the vacancies, a stronger material with a much greater energy absorbing capability can be produced. This Metal foams are materials which display a unique has the effect of increasing the stability of the cell walls, combination of physical and mechanical properties. Their reducing the likelihood of their buckling under loading. light weight, high specific stiffness, high strength to weight While doing this obviously increases the density of the ratios, and greatly increased energy absorbing capabilities material, the resulting strength to density ratios are higher make them ideal candidates for use in the automotive and than those reported in the literature for foams processed using aerospace industries.1) They have been shown to experience similar materials and different techniques. The physical fatigue degradation in both tension and compression.2,3) properties of these materials can be altered by changing the Under compression, localized failure begins preferentially at size and wall thickness of the hollow spheres. By using larger cells within the assembly, which eventually leads to preformed hollow spheres of known dimensions, the proper- the formation of collapse bands.2,3) ties of the foam become more uniform and predictable. The performance of existing foams has not been promising These processes allow both similar and dissimilar matrix due to strong variations in their cell structure.2–4) Most and hollow sphere materials to be used in manufacturing of commercially available cellular metals do not achieve the composite metal foams. In the casting process, dissimilar properties predicted from the scaling relations that connect materials must be used; specifically the matrix must be of a the mechanical behavior of the foam to the bulk material they lower melting temperature material than that of the hollow are produced from.5–7) This can be partially attributed to spheres. In the PM process, similar materials can be used morphological defects in the structure such as missing cell because the processing temperature is below the melting walls, wiggles in the cell wall, etc.7) To gain the full temperature of the components. Additionally, similar materi- advantages of these lightweight materials, these defects must als may lead to better bonding strength between the matrix be eliminated.8) and the sphere walls. One technique that has been previously explored is to use preformed hollow spheres to form the cells of the material. 2. Materials, Processing and Characterization Two such hollow sphere materials have been created and studied by Georgia Tech and Fraunhofer.9–11) These materials 2.1 Material and equipment exhibited plateau stresses of 5 and 23 MPa respectively, with The hollow spheres used in these studies are produced by energy absorbing capabilities of 2 and 10 MJ/m3 respective- Fraunhofer in Dresden Germany, using a PM technique.9,11) ly, up to 50% strain. The composition of the low carbon (LC) steel spheres is We have produced a new type of closed cell foams by <0:002% oxygen, <0:007% carbon, and the balance iron. filling the vacancies around a random dense collection of Two sets of LC steel hollow spheres were produced with preformed hollow spheres with a solid matrix material, either nominal outer diameters of 3.7 and 1.4 mm with wall thick- through casting12,13) or through powder metallurgy (PM)12,14) nesses of 200 and 50 mm, respectively. The porosity of the with the aim of increasing their energy absorption and sphere walls is approximately 5% for both sizes. The 316L strength. The studies presented in this paper have shown that stainless steel hollow spheres have a nominal outside diameter of 2 mm with a wall thickness of 100 mm. The *1Graduate Student, North Carolina State University porosity of the sphere walls ranged from 6–15%. *2Undergraduate Student, North Carolina State University The solid matrix in these studies is produced by two Processing and Characterization of a New Composite Metal Foam 2149

Table 1 Matrix alloying elements in composite foams (mass%). The mold and spheres are preheated in the furnace at 700C. Aluminum 356 Casting Alloy This helps prevent premature solidification of the aluminum when it is cast into the mold. The aluminum is simulta- Element Si Fe Mg Mn Cu Ti neously melted in a graphite crucible in the furnace. The mass% 7.01 0.5 0.39 0.28 0.11 0.09 aluminum is cast into the mold and allowed to air cool before Element Zn Cr Al — — — removal. More detail on the casting procedure is available mass% 0.06 0.02 Bal — — — elsewhere.13) Ancorsteel-1000C steel powder 2.2.2 Powder Metallurgy Element C P S Si O Ni For the PM process, hollow spheres are placed inside the mass% 0.003 0.006 0.007 0.002 0.005 0.03 mold and vibrated to arrange them into a denser packing Element N Mo Mn Cu Cr Fe configuration. Vibration was done using an APS Dynamics mass% <0:001 0.02 0.1 0.05 0.02 Bal. model 113 shaker and a APS model 114 amplifier with a 316L stainless steel powder General Radio 1310-B frequency generator. This is followed Element Cr Si Mn Ni Mo C by adding the powder and further vibration to completely fill the spaces between the spheres. The vibration time was 30 mass% 17 0.9 0.2 13 2.2 0.03 min at 15–20 Hz. Element O Fe — — — — The sintering cycle consisted of heating the sample at mass% 0.3 Bal — — — — 10C/min up to 850C, a 30 min soak at 850C, heating at 5C/min up to 1200C, a 45 min soak at 1200C and cooling at 20C/min. A duplex cycle has been selected to achieve different methods. In the casting method, aluminum 356 higher mechanical properties.17) The lower temperature step casting alloy was used. In the PM method, low carbon causes a reduction of remaining oxides and removal of Ancorsteel-1000C from ARC Metals was used for the LC organic impurities and helps bring the mold to thermal steel composite foam. An additional 0.8% graphite was equilibrium to avoid gradients in properties. Surface trans- mixed in with the steel powder for strengthening15) and to port effects are most prevalent at lower temperature, so the control shrinkage of the matrix during sintering.16) The particle bonds are strengthened without densification.18) At powder used for the stainless steel foam is 316L stainless higher temperatures, strength is increased greatly as a result steel from ANCOR Specialties. No additions were made to of the higher sintering rate due to greater atomic motion. For the stainless steel powder. The compositions of the matrix both temperatures, rapid increases in strength have been materials are shown in Table 1. reported for soaking times up to 30 minutes.15) The molds for producing these materials had to be large enough to maintain a minimum of 8–10 spheres per edge to 2.3 Characterization eliminate any edge effects so as to produce a meaningful Samples for SEM and optical observation were prepared representation of a bulk material.2) Other considerations were using Buehler Automet 2 Power Head grinding and polishing simplicity of manufacture, ease of assembly and ease of part stations. Grinding was done with 240, 400, and 600 grit removal. The mold for the casting method is a gravity fed paper, using a 2.3 kg load and a wheel speed of 60 RPM for design made of carbon steel, with a mold cavity of 121 mm Â the cast foam and 90 RPM for the PM foam. Polishing was 144 mm Â 54 mm. The mold for the PM method was made done with 9, 3, and 1 mm diamond suspension polish and from 304 stainless steel, with a mold cavity of 51 mm Â finished with 0.05 mm alumina paste, using a 1.8 kg load and a 51 mm Â 89 mm. Prior to processing, the surfaces of the wheel speed of 60 RPM for the cast foam and 150 RPM for molds are cleaned and then coated with a boron nitride mold the PM foam. The samples were cleaned in an ultrasonic release to prevent bonding between the sample and the mold cleaner between each polishing step. and to facilitate removal from the mold. To avoid the build- Optical microscopy was performed using a Buhler Unitron up of oxides from inside the mold, non co-linear holes were 9279 optical microscope equipped with a Hitachi KP-M1 drilled into the spacer and mold cap. This allows venting of CCD black and white digital camera, and Omnimet image the atmosphere and any outgassing inside the mold, but holds grabber software. the powder in place. SEM Images were taken with a Hitachi S-3200N environ- The furnace used for casting is 3300 series high temper- mental SEM equipped with EDX to determine the micro- ature laboratory furnace from CM Furnaces, with molydisi- structure and chemical composition of the samples. licide heating elements capable of reaching 1700C. Monotonic compression testing was performed using an The hot press used in this experiment is a Centorr 600- MTS 810 with a 980 kN load cell and a crosshead speed of 4X6W4-26HP vacuum hot press with a vacuum operating 1.25 mm/min. level of 2 Â 10À3 Pa, tungsten heating elements capable of Microhardness testing was performed using a Buehler reaching 2600C and a hydraulic ram, which can apply a Micromet microhardness tester. For the cast foam and the LC pressure of 135 MPa through a 25.4 mm diameter rod. steel foam, a 20 g load was used. For the stainless steel foam, a 50 g load was used. 2.2 Processing 2.2.1 Casting 3. Results and Discussion For casting, the spheres are placed inside the mold and a steel screen is fixed over the spheres to hold them in place. Figure 1 shows digital images from the sectioned samples 2150 A. Rabiei, L. Vendra, N. Reese, N. Young and B. P. Neville

Fig. 1 Cut sections of the composite Al-steel cast foam, 3.7 mm and 1.4 mm composite steel-steel PM foam (from left to right).

3.1 Cast foam The cast sample had a density of 2.4 g/cm3 (42% relative density). During monotonic compression, this foam reached an average plateau stress of 67 MPa up to 50% strain, before it began densification around 50% strain. Optical and SEM observation was used to determine how well the spheres had bonded with the matrix material. It was determined that the matrix had nearly filled all of the vacancies between the spheres (Fig. 3(a)). It has been calculated that there is less than 1% void space due to the aluminum failing to completely fill the interstitial spaces between the point contacts of the spheres.13) SEM images showed different phases present in the matrix. SEM-EDX compositional analysis was performed Fig. 2 Composite metal foam comparison under monotonic compression. to differentiate the phases. The sphere walls were verified to be nearly 100% iron. The matrix was found to be nearly 98% aluminum, with the remainder being silicon and trace of the Al-LC cast foam as well as the 3.7 mm LC foam and elements. The compositions of the other phases are shown 1.4 mm LC foam produced by PM. in Table 3. Figures 3(b) and 3(c) show the other phases that Monotonic compression testing of all materials demon- are present within the matrix. strated the typical behavior of an elastic-plastic foam under Microhardness testing was performed on all various compression.4) There is an initial linear elastic region, which components and phases present in the cast samples. Accurate is followed by an extended region of deformation at a readings could not be taken for the light and dark needle relatively constant level of stress. Unlike most foams, shaped phases due to their small size. These results are shown however, these materials do not exhibit a level plateau stress. in Table 3. Two different phases were identified as present The material densifies at a slowly increasing rate and there is within the matrix- the light gray phase and the dark gray no distinct point at which full densification occurs. For this phase. The light gray phase, a ternary alloy of Al with Si and paper, the plateau stress has been defined as the average stress Fe, estimated to be Al4FeSi from the Al-Fe-Si ternary phase from 0.2% (yield point) to 50% strain. All materials reached a diagram19) was found in two different shapes-plates and minimum of 50% strain and had not yet reached full needles. Binary Al-Si alloys with increased iron content form densification at that point. Figure 2 shows the stress-strain hard, brittle plates of the compound -AlFeSi and with curves of various composite foams under monotonic com- controlled iron content form the finer -AlFeSi compound.20) pression. After 50% strain, the foams begin to approach The plate shaped light gray phase found concentrated around densification as the hollow spheres are completely collapsed the sphere walls as shown in Fig. 3(c) is the hard -AlFeSi and the material begins to behave like a bulk material. The compound (Table 3) whose formation was facilitated by the physical properties of the foams processed using both diffusion of iron from the sphere walls into the aluminum techniques are shown in Table 2. matrix. This is also the reason why the plate shaped light gray Processing and Characterization of a New Composite Metal Foam 2151

Table 2 Comparison of physical and mechanical properties of composite metal foams.

Sample Property Cast Foam Low Carbon Steel PM Foam Stainless Steel PM Foam Dimensions 36 Â 40 Â 53 44:0 Â 44:0 Â 55:024:0 Â 24:0 Â 38:620:3 Â 20:3 Â 34:9 (L Â W Â H, mm) Sphere OD (mm) 3.7 3.7 1.4 2.0 Sphere Wall 0.2 0.2 0.05 0.1 Thickness (mm) Sample Density (g/cm3) 2.4 3.2 2.6 2.9 Relative Density (%) 42 40.7 34.2 36.8 Plateau Stress (MPa) 67 42.3 76 136 Densiﬁcation 50 55 50 50 Strain (%) Strength/Density 28 13.2 29.8 46.9 Ratio (MPa/(gÁcmÀ3)) Energy Absorbed 32.3 21 37.6 68 up to 50% Strain (MJ/m3)

Table 3 Compositional analysis and corresponding Vickers Hardness of phases in aluminum matrix (atomic %) in cast sample.

Feature Al Si Fe HV Spheres — — 100 69.7 Aluminum matrix 97.9 2.6 Trace 45.1 Light grey plate 64.95 9.27 25.78 732.5 shapes phase Light grey needle 66.46 19.35 14.19 — shaped phase Dark grey phase 3.02 96.87 0.07 —

the matrix to the sphere-matrix interface facilitates the precipitation of plate shaped light gray phase at the sphere- matrix interface. The needle shaped light gray phase within the matrix away from the sphere walls (Fig. 3(b)) is the finer and possibly softer -AlFeSi compound whose formation was facilitated by the controlled iron content, and also by presence of manganese20) and silicon which comprises the composition of Al 356 alloy. Slow solidification of a pure Al-Si alloy produces a very coarse microstructure in which the eutectic comprises of large plates or needles of Si in the continuous matrix, formed due to precipitation of silicon.20) The dark gray phase found in the matrix is primarily this eutectic of approximately 97% Si and 3% Al composition, formed due to the precipitation of Si from the solid solution at elevated temperatures. The flake- like shape of the dark gray phase is attributed to the highly anisotropic crystal growth in diamond cubic systems like silicon. The composition and hardness values of all the phases in the aluminum matrix are shown in Table 3. Microhardness test results show that the diffusion of iron Fig. 3 SEM image showing aluminum matrix in a cast sample (a) between from the spheres into the aluminum takes place leading to the two spheres, (b) Dark grey (DG) and light grey (LG) needle shaped phase, precipitation of a hard plate-shaped light gray phase, and (c) Light grey plate shaped phase in Al-LC steel cast composite foam. subsequently strengthening the material.

3.2 PM Foam phase is always present only in the immediate vicinity of the Three different sets of foams have been produced using the spheres, as its formation requires higher iron content along PM technique. Two have been produced using LC steel with aluminum and silicon available only at the sphere- powder with 3.7 and 1.4 mm low carbon steel spheres with matrix interface. The diffusion of iron from sphere walls into similar matrix, and one 316L stainless steel powder and the aluminum matrix as well as the diffusion of silicon from 2.0 mm stainless steel spheres. The first PM set of samples, 2152 A. Rabiei, L. Vendra, N. Reese, N. Young and B. P. Neville

Fig. 4 Photographs of the PM stainless steel composite foam before (left) Fig. 5 SEM image of 1.4 mm hollow sphere LC steel composite foam with and after (right) monotonic compression of 60%. dashed line showing the wall thickness of one sphere (note that the cut is not through the center of the spheres). produced with the 3.7 mm LC steel spheres, had a density of 3.0 g/cm3 (38% relative density). During monotonic compression, this foam reached a plateau stress of 42.3 MPa before beginning densification around 55% strain. The second set of LC steel PM samples, with the 1.4 mm spheres, had a density of 2.55 g/cm3 (34% relative density) and reached a plateau stress of 76 MPa during monotonic compression before it began densification around 50% strain. The 316L stainless steel PM sample, produced with 2.0 mm spheres, had a density of 2.9 g/cm3 (37% relative density) and reached a plateau stress of 136 MPa up to 50% strain and began densification around 50% strain. Figure 4 shows the stainless steel foam before and after compression testing with 60% strain. These plateau stresses compare favorably to other hollow sphere foams with similar materials, which exhibit plateau stresses of 5 and 23 MPa.9–11) SEM observation showed that the powder was able to completely fill up all spaces between spheres and SEM Fig. 6 SEM image of 2.0 mm 316L stainless steel composite foam showing images show that the only distinction between the sphere porosity difference in sphere walls and matrix, with dashed line showing walls and the sintered matrix is the visible density of the wall thickness of one sphere. porosities in the matrix (Figs. 5, 6). Microhardness was performed on the PM foams as well. Results are shown in Table 4. Due to the porosity of the Table 4 Vickers Hardness of different regions in the LC and stainless steel matrix, there was difficulty in attaining accurate readings. PM foams. Some indentations were greatly distorted and others were on Stainless edges of pores. Only readings where the indentation was able Low Carbon Low Carbon Stainless Region Steel to be read accurately were used to calculate the average. Steel Matrix Steel Spheres Steel Matrix Spheres Although the matrix is made of the same material as the spheres in both cases, the spheres averaged about 40 HV HV 49.6 88.9 142.0 186.2 harder than the matrix. The lower hardness of the matrix is due to the greater amount of porosity within the matrix. The energy absorbing capabilities of the new foams are Due to the sintering temperature and vacuum level, there superior to those of other hollow sphere foams. Most of the was some concern that the chromium in the stainless steel new foams were able to absorb almost 2–4 times more energy powder would evaporate, reducing the corrosion resistance than the other foams,12) with the stainless steel PM foam capability of the matrix for the stainless steel foams.15) SEM having an energy absorbing capability almost 7 times greater EDX was performed on the sintered matrix and it was found than foams made from similar materials through different that the composition of the stainless steel was within the techniques. Because the foams begin densification at differ- nominal composition of the powder as given by the ent strains, it was decided to compare them all at 50% strain. manufacturer (Table 1). Processing and Characterization of a New Composite Metal Foam 2153

4. Summary A. G. Evans: Acta Mater. 45 (1997) 5245–5259. 5) M. F. Ashby, A. G. Evans, N. A. Fleck, L. J. Gibson, J. W. Hutchinson and H. N. G. Wadley: Metal Foams: A Design Guide, (Butterworth- The strength and energy absorbing capabilities of hollow Heinemann, 2000). sphere foams have been greatly increased by ﬁlling in the 6) A. G. Evans, J. W. Hutchinson and M. F. Ashby: Progress in Materials vacancies around the hollow spheres with a solid matrix Science 43 (1998) 171–221. material. By using two diﬀerent techniques, casting and 7) L. J. Gibson and M. F. Ashby: Cellular Solids, Structures and powder metallurgy, both similar and dissimilar materials Properties, 2nd, (Cambridge University Press, Cambridge, UK, 1997). 8) U. Ramamurty and A. Paul: Acta Materialia 52 (2004) 869–876. can be used for the matrix and hollow spheres. Further study 9) G. Stephani, D. Kupp, T. D. Claar, G. Stephani and U. Waag: 2001 on mechanical properties of our composite metal foams, International Conference on Powder Metallurgy and Particulate especially under cyclic loading, as well as their failure Materials, 50–58, Fraunhofer Institute For Manufacturing And Ad- modeling, sound and heat management is being investigated. vanced Materials., 2001. 10) T. J. Lim, B. Smith and D. L. McDowell: Acta Mater. 50 (2002) 2867– 2879. Acknowledgements 11) O. Andersen, U. Waag, L. Schneider, G. Stephani and B. Kieback: Advanced Engineering Materials 2 (2000) 192–195. The authors would like to acknowledge the support from 12) A. Rabiei, A. T. O’Neill and B. P. Neville: MRS Fall 2004 Proceedings NSF-DMII for funding this research through CAREER award 841 (2005) 517–526. # 0238929. 13) A. Rabiei and A. T. O’Neill: Mater. Sci. Eng. A 404 (2005) 159–164. 14) B. P. Neville and A. Rabiei: Mater. Sci. Eng. A, 2006. 15) ASM Metals Handbook, 9th, American Society for Metals, 1984. 16) N. Dautzenberg and J. Hewing: Powder Metallurgy International, 9, REFERENCES 1977. 17) L. Forss: Perspectives in Powder Metallurgy, (Plenum Press, New 1) J. Banhart and W. Brinkers: Journal of Material Science Letters 18 York, 1968). (1999) 617–619. 18) R. M. German: Powder Metallurgy of Iron and Steel, (John Wiley and 2) Y. Sugimura, A. Rabiei, A. G. Evans, A. M. Harte and N. A. Fleck: Sons, New York, 1998). Mater. Sci. Eng. A 269 (1999) 38–48. 19) P. Villars, A. Prince and H. Okamoto: Handbook of Ternary Alloy 3) A. Rabiei, A. G. Evans and J. W. Hutchinson: Metall. and Mater. Trans. Phase Diagrams, (1995). 31A (2000) 1129–1136. 20) I. J. Polmear: Light Alloys-Metallurgy of the Light Metals, 3, (1995). 4) Y. Sugimura, J. Meyer, M. Y. He, H. Bart-Smith, J. Grenstedt and