Bruno Pinto Lopes Desenvolvimento de pre ursores para o fabri o de
espumas metáli as
Development of pre ursors to produ e metalli
Bruno Pinto Lopes Desenvolvimento de pre ursores para o fabri o de
espumas metáli as
Development of pre ursors to produ e metalli
foams
DissertaçãoapresentadaàUniversidadedeAveiropara umprimentodos
req-uisitosne essáriosà obtençãodo graude Mestre emEngenharia Me âni a,
realizada sob orientação ientí a de Doutora Isabel Maria Alexandrino
Duarte, Investigadora Auxiliar do Departamento de Engenharia Me âni a
da Universidade de Aveiro e de Prof. Doutora Móni a Sandra Abrantes
De Oliveira Correia, Professora Auxiliar do Departamento de Engenharia
Presidente/ President Prof. Doutor Rui Pedro Ramos Cardoso
ProfessorAuxiliardaUniversidadedeAveiro
Vogais /Committee Prof. Doutor José Maria daFonte Ferreira
ProfessorAsso iado /agreg. daUniversidadedeAveiro
Doutora IsabelMaria Alexandrino Duarte
InvestigadoraAuxiliardaUniversidadedeAveiro(orientador)
Prof. Doutora Móni aSandra Abrantes De OliveiraCorreia
that had the help of several people and businesses. Thus I want to pay a
spe ialthanks to my advisors: Dr. Isabel Duarte and Dr. Móni a Oliveira
forsuggestions, helpand tirelesssupportprovided during the work.
An a knowledgement to ompanies Chemetall (Frankfurt, Germany) and
Alpo o (Nottingham, UK) for supplying the powders of titanium hydride
and aluminiumalloysrespe tively. To MJAmaral ompanyinthe person of
Mr. Manuel Vide to the support on manufa turing the mould to produ e
the pre ursormaterial.
Ialso want tothank ProfessorDr. José Maria FerreirafromDepartment of
Cerami s and Glass ofthe Universityof Aveiro forthe helpand availability.
Also, I would like to thank the laboratory te hni ians Célia Miranda and
Ana Ribeiro also from Department of Cerami s and Glass of University of
Aveiroforsupportin the useofequipment for arrying out thework.
Finally I would like to thank Inês for all her support and help provided
during this dissertation.
Abstra t The metal foams produ ed by powder metallurgy have been in reasingly
used in various appli ations. Currently, these materials are being used in
lightweightstru tures,energyabsorptionandsoundmuingin ars,trains,
tramsand air rafts. Themain obje tiveofthis thesis was the development
of quality pre ursormaterialforthe manufa tureofAluminiumalloyfoams.
Forthat, ommer ialpowders ofthree typesofAluminium alloys(Al-alloy)
and two types of Titanium Hydride (TiH
2
) of dierent hemi al omposi-tionsand sizeswere tested. To attainthe latter, the workwasdivided intofourphases: (1)Evaluationofthe powderproperties,in luding thestudy of
oxidation and thermal de omposition rea tionof TiH
2
powders by thermal analysis, omplemented by X-ray dira tion (XRD), s anning ele tronmi- ros opy (SEM)and energydispersivespe tros opy X-ray (EDX),and also
the thermal behaviour of Al alloys by hot mi ros opy te hnique. (2)
Op-timization of the parameters in the manufa turing pro ess of the pre ursor
material. (3) Evaluation of the ability of the pre ursormaterials developed
in the previous stepto produ e Alalloy foams withgood quality. (4)
Eval-uation ofthe propertiesof theobtainedfoams inthis work,su has density,
ompressive strength and energy absorption apa ity. As a result of this
work, only two of the pre ursor materials developed during the exe ution
of this workhave demonstratedto be able tomanufa ture aluminium alloy
foamswithquality. Thesepre ursorsmaterialswereobtainedusingpowders
ofthesametypeofAl-alloy(Al88.45%and11.19%Sili on)withanaverage
diameter of 16
µ
m, only diering in the type of used TiH2
powders (6.89µ
m and 15.39µ
m). Both TiH2
powders were previously thermally treated at 480◦
C during 180 minutes. From these studies, it an on luded that
it ispossible toobtain pre ursors with densitiesabove80% of the
theoreti- al density by applying a ombinationof old and a hot pressing at 400
◦
C
using 200bar. The pre- ompa ted material at old temperature must be
heatedto400
◦
Candkeptatthistemperatureduring40minutes,beforehot
pressing. TheAl-alloy foams obtained using TiH
2
powder with an average diameterof6.89µ
mpresenteddensitiesbetween652.28and527.44kg/m3
.
The obtained foamsusing TiH
2
powder withan average diameter of 15.39µ
m haveshown lowerdensity values, ie. between 487.8and 421.29 kg/m3
.
Fromthe evaluationof the me hani alproperties,itwas found thathigher
density foams havebetterme hani alproperties, namelyYoung's modulus,
stress leveland energy absorption apa ity. For example, Young's modulus
forthe higher density foamslies between 106.30and 147.54 MPa, whereas
Resumo Asespumasmetáli asproduzidaspormetalurgiadepóstêmvindoaser ada
vez mais utilizadas nas mais diversas apli ações. A tualmenteestes
mate-riais estão aser utilizadosem estruturas ultraleves, de absorção de energia
e de amorte imento sonoro em veí ulos automóveis, omboios,elé tri ose
aeronaves. Oprin ipalobje tivo desta dissertaçãoprendeu-se omo
desen-volvimento dematerialpre ursorde qualidadepara ofabri ode espumasde
ligasde Alumínio. Parao efeitoforamtestadospós omer iaisde trêstipos
deligasdeAlumínio(ligadeAl)edoistiposdeHidretodeTitânio(TiH
2
)de diferentes omposições quími ase granulometrias. Para atingiresteobje -tivo, o trabalho dividiu-se em quatro fases: (1)Avaliação das propriedades
dos pós, in luindoo estudo daoxidação eda rea çãode de omposição
tér-mi a dos pós de TiH
2
através de análises térmi as, omplementadas om difra ção de raio-X(DRX), mi ros opia ele tróni ade varrimento(SEM)eespe tros opiade energia dispersiva de raio-X (EDX), e ainda o estudo do
omportamentotérmi odasligasde Alatravésdaté ni ade mi ros opiaa
quente. (2)Optimizaçãodosparâmetrosno pro essodefabri odo material
pre ursor. (3)Avaliaçãoda apa idadedosmateriaispre ursores
desenvolvi-dosnaetapaanteriorparaproduzirespumasdeligasdeAlde qualidade. (4)
Avaliação das propriedades das espumas obtidas neste trabalho,
nomeada-mente a densidade, a resistên iaà ompressão e a apa idadede absorção
de energia. Comoresultado deste trabalho, apenas dois dosmateriais
pre- ursoresdesenvolvidosdurante aexe uçãodestetrabalhodemonstraramter
apa idade de fabri ar espumas de ligas de Alumínio de qualidade. Estes
materiaispre ursoresforamobtidosusandopósdo mesmotipodeligadeAl
(88.45% de Ale11.19% de Silí io) de diâmetromédio de 16
µ
m, diferindo apenas no tipo de pós de TiH2
usados (6.89µ
m e 15.39µ
m). Ambos os pósdeTiH2
forampreviamentetratadostermi amentea480◦
Cdurante180
minutos. Dosestudosrealizados, on lui-sequeépossívelobterpre ursores
om densidades superiores a 80% da densidade teóri a através da
ombi-nação de uma prensagem a frioe de uma prensagem a quente a400
◦
C de
200bar. O pré- ompa tado a frio deve ser aque ido até 400
◦
C e estar a
esta temperatura durante 40 minutos, antes da prensagem a quente. As
espumas de ligas de Alobtidas usando pós de TiH
2
de diâmetro médio de 6.89µ
mapresentaramdensidadesentre652.28e 527.44kg/m3
. Asobtidas
usando pós de TiH
2
de diâmetromédiode 15.39µ
mapresentaram valores de densidade inferiores, ou seja entre 487.8 e 421.29 kg/m3
. Da avaliação
daspropriedadesme âni as,veri ou-sequeasespumasde maiordensidade
apresentam valores de propriedadesme âni assuperiores,nomeadamenteo
módulode Young, tensãode patamare apa idadede absorçãode energia.
List of Tables iii
List of Figures iv
Nomen lature ix
1 Introdu tion 1
2 State of the Art 3
2.1 Introdu tion . . . 3
2.2 Classesof metalli foams . . . 4
2.3 Manufa turingPro esses . . . 5
2.3.1 Produ tion offoams frommetalli melts . . . 5
2.3.1.1 Dire tfoaming methods . . . 5
2.3.1.2 Solid-gas eute ti solidi ation . . . 6
2.3.1.3 Investment asting . . . 8
2.3.1.4 Synta ti foams using llermaterials. . . 8
2.3.2 Foams madefrom metalpowders . . . 9
2.3.2.1 Powder metallurgy . . . 9
2.3.2.2 Foaming of slurries. . . 11
2.3.2.3 Gas entrapment . . . 11
2.3.2.4 Other Te hniques . . . 11
2.3.3 Produ tion byDepositionTe hniques . . . 12
2.4 Properties . . . 12
2.4.1 Density . . . 15
2.4.2 Stru tural properties . . . 15
2.4.3 Me hani al Properties . . . 17
2.4.3.1 Stress-strainbehaviourunder ompression. . . 17
Linear elasti region . . . 18
Plateau region. . . 20
Densi ation region. . . 21
2.4.3.2 Energy Absorptionunder ompression . . . 22
2.4.4 Thermal properties . . . 25
2.4.5 A ousti properties . . . 25
2.4.6 Ele tri al properties . . . 28
2.5 Appli ations . . . 29
2.5.4 Railwayindustry . . . 35
2.5.5 Building industry . . . 35
3 Experimental Pro edures 37 3.1 Materials . . . 37
3.1.1 Powder spe i ations. . . 37
3.1.2 Preparation ofblowing agents powders . . . 38
3.1.3 Preparation ofpre ursor materials . . . 39
3.1.4 Preparation ofAl-alloys foams . . . 40
3.2 Chara terization Methods . . . 41
3.2.1 Powder analysis. . . 41
3.2.1.1 Density . . . 41
3.2.1.2 Parti le size analysis . . . 42
3.2.1.3 Thermal analysis . . . 43
3.2.1.4 X-ray dira tionanalysis . . . 44
3.2.1.5 Hot-stage mi ros opy analysis . . . 44
3.2.2 Mi ros opi analysis . . . 45
3.2.2.1 Light mi ros opy(LM)analysis . . . 45
3.2.2.2 S anningele tron mi ros opy(SEM)analysis . . . 46
3.2.3 Me hani al analysis . . . 46
4 Results and Dis ussion 49 4.1 Powders . . . 49
4.1.1 Density . . . 49
4.1.2 Parti lesize . . . 50
4.1.3 Morphology analysis . . . 51
4.1.4 Thermal de omposition and oxidation behaviour of the Titanium Hydridepowders . . . 53
4.1.4.1 Thermal behaviourof TiH
2
powdersinAir atmosphere . 55 4.1.4.2 ThermalbehaviourofTiH2
powdersinOxygenatmosphere 59 4.1.4.3 X-ray dira tioninvestigations . . . 614.1.5 Thermal Behaviour oftheAl-alloypowders . . . 70
4.2 Produ tion ofpre ursor material . . . 74
4.3 Produ tion ofAl-alloys foams . . . 81
4.3.1 Preliminarytests . . . 81
4.4 Properties ofAl-alloyfoams . . . 84
4.4.1 Preparation . . . 84 4.4.2 Density . . . 86 4.4.3 Compression behaviour . . . 87 4.4.4 Energy absorption . . . 91 5 Con lusions 95 6 Future works 97
2.1 Ranges forproperties of ommer ial foams . . . 14
3.1 Supplierspe i ation ofthe Al-alloy powders . . . 37
3.2 Supplierspe i ation ofthe TitaniumHydridepowders . . . 38
3.3 Manufa turingparameters usedto produ e Al-alloyfoams . . . 47
4.1 DensityoftheAl-alloyand TiH
2
powders. . . 494.2 Parti lesize analysisbylaserdira tion. . . 50
4.3 Optimum ompa tion onditions . . . 77
4.4 Foamable pre ursormaterials . . . 78
4.5 Densities andrelative densityof Al-alloyfoams . . . 86
4.6 Inuen eof the foam omposition and densityonme hani al results . . . 89
4.7 Energy absorption apability and absorption energy per unit volume at 50%strain andat densi ation strain. . . 91
2.1 Metalli foams. a) open- ell stru ture. b) losed- ell stru ture . . . 4
2.2 Manufa turingpro essesof metalli foams . . . 5
2.3 Dire t foaming ofmelts bygasinje tion . . . 6
2.4 Dire t foaming meltsby addingblowing agents . . . 7
2.5 Solid-gas eute ti solidi ation . . . 7
2.6 Investment astingpro ess . . . 8
2.7 Syntati foams using llermetal. . . 9
2.8 Powder metallurgy pro ess. . . 10
2.9 Gas entrapment te hnique . . . 11
2.10 Properties of ellular and denseengineering materials . . . 13
2.11 The range of ell size and relative density for thedierent metalli foam manufa turing methods . . . 13
2.12 Young modulusfor dierent foam . . . 15
2.13 List ofparameters todes ribethestru ture of metalli foam . . . 16
2.14 Closed- ell Aluminiumfoam . . . 16
2.15 Closed- ell foamwithgradient density . . . 17
2.16 Compressive stress-strain urves of a) elastomeri , b) elasti -plasti and )elasti -brittle foam. . . 18
2.17 Ee t offoam densityon ompressive stress-strain urves . . . 18
2.18 Deformation me hanisms of foams: a) open- ell foam, sequentially ell wall bending, ell wall axialdeformation and uidowbetween ells and b) losed- ellfoams,sequentially ell wallbendingand ontra tion, mem-branestret hinganden losed gaspressure . . . 19
2.19 Cubi unit ell models of open- ell and losed- ell foams developed by Gibsonand Ashby . . . 19
2.20 Variationof Young's modulewithdensity . . . 21
2.21 Stress-strain urvesfor both an elasti solidand a foam showing the dif-feren e inenergy perunit volume absorbed at thesame stresslevel . . . . 22
2.22 Stress-strain urvesof foamsfor three dierent density . . . 23
2.23 Comparison oftheideal absorberanda realabsorber . . . 24
2.24 Maximum stress o urring when a given deformation energy is absorbed byAluminiumfoams withvariousdensities . . . 24
2.25 Thermal ondu tivityplottedagainstvolumetri spe i heatfor urrently available metalli foams . . . 26
2.26 Sound absorption oe ient of various typesof ellular Alfoamsin om-parison withber lasssoundabsorbingmaterial . . . 27
thesample, 20 mmair gapbetween sample andrigid wall) . . . 27
2.28 Dampingfa tor andvibration behaviorfor Alulight foam. . . 28
2.29 Thepowerlawdependen eofnormalizedele tri ondu tivityofAl-foams on the relative density for Alulight foam . . . 29
2.30 Magneti eld shielding ee tiveness as a fun tion of frequen y for Alu-minium foam(Alulight) andsteelsamples ofthesame weight and a mas-sive Al sheet of the same thi kness (t=8.5 mm) asthe Aluminium foam. Sample size: 140x140 xt,density offoam500kgm
−3
. . . 302.31 Appli ations of ellular metals grouped a ordingto the degreeof open-ness needed andwhether theappli ation ismore fun tionalor stru tural 31 2.32 Appli ations eldsfor losed- ell metalli foams . . . 31
2.33 Automotive omponents . . . 32
2.34 Implementation ofmetalli foamsina ar . . . 33
2.35 Realautomotive appli ations for Al-alloyfoams . . . 33
2.36 Arianero ket one: prototypeassembly. 3DfoamedAFS6061+AlCu6Cu6 34 2.37 Sandwi h panelmetalfoams . . . 34
2.38 Building appli ationsfor Al-foams . . . 35
3.1 TitaniumHydride powders: as-re eived (left side) andtreated (right side) 38 3.2 Plasti bottles ontaining thedierent prepared powdersmixture . . . 39
3.3 Tumbling mixer usedinthis resear h to mixthepowders. . . 39
3.4 Overview of the system in the laboratory to manufa ture the pre ursor materials. a) Uni-axial pressing devi e with thedie and the heating sys-tem. b) Pressingdevi e usedfor thepressingthepowder mixture . . . 40
3.5 Furna e usedfor themanufa ture ofAl-alloyfoams . . . 41
3.6 Py nometer usedduring analysis . . . 42
3.7 LS 230 LaserDira tion Parti le Size AnalyzefromBe kman Coulter . . 43
3.8 (a) Thermal analyser LabsysTM TG-DTA16. (b) Simultaneous Thermal Analyser STA 449 CNetzs yh . . . 44
3.9 Hot-StageMi ros opy(Leitz, model2A) . . . 45
3.10 light mi ros opeNikon E lipse LV150 . . . 45
3.11 S anning Ele tronMi ros ope (HR-FESEM Hita hi SU-70) . . . 46
3.12 Universaltestingma hine(Shimadzu AG-50kNG) . . . 47
3.13 Al-alloyfoams tested . . . 47
4.1 SEMmi rographofAl-alloypowdersatdierentmagni ations. Leftside: overview. Right side: parti lesdetails. . . 52
4.2 SEMmi rograph of as-re eived TiH
2
powdersat dierent magni ations. Leftside: overview. Right side: parti le details. . . 534.3 Thermo analysisofthe TiH
2
powder samples performed at 10k/min. . . . 544.4 TG/DTG urves measured on as-re eived and treated TiH
2
powders in Air at10k/min. . . 564.5 TG urves measured on as-re eived and treated TiH
2
powders in air at 10K/min. . . 584.8 XRD-measurements on TiH
2
powder samples as-re eived and treated at 480◦
Cduring3 hours. . . 62
4.9 Referen e spe tra for TiH
2
,TiO2
,TiO3
and TiN0
.3
.. . . 63 4.10 XRD-measurements on as-re eived TiH2
powder samples and aftersub-mitted inAir and Oxygenatmosphere. . . 64
4.11 SEM mi rographs of surfa e of a parti le of TiH
2
powder submitted at dierent onditions . . . 654.12 SEM mi rographs of surfa e of a parti le of TiH
2
powder submitted at dierent onditions . . . 664.13 EDX spe trum ofTiH-1 powder . . . 67
4.14 EDX spe trumofTiH-2 powder . . . 68
4.15 EDXspe trum ofTiH-2powderaftersubmittedatairatmosphere inFig.
4.12b . . . 69
4.16 TiH
2
oxidised at dierent onditions. The olour of the powder depends on the thi kness ofthe Oxygenlayer. a ) as-re eveidb) treated at 480◦
C
for 3h )submitted inOxygen atmosphere d)submitted inAiratmosphere. 69
4.17 Hot-stagemi ros opyimagesofasampleofAl-Aalloypowderatdierent
temperatures(before, duringand aftermelting) . . . 71
4.18 Hot-stage mi ros opy images of a sample of Al-Balloypowder (32.5
µm
) at dierent temperatures(before, duringand aftermelting) . . . 724.19 Hot-stagemi ros opyimagesof asampleofAl-Calloypowder (240.9
µm
) at dierent temperatures(before, duringand aftermelting) . . . 734.20 S hemati representation of experimental tests to adjust the ompa tion
parameters usinghot ompa tionstage (a)and a ombinationof oldand
hot ompa tion stages(b) . . . 74
4.21 Ee t of ompa tion parameters on thedensityof thepre ursormaterial . 75
4.22 Low-quality ofpre ursormaterials whi h wasobtained byhot pressingat
dierent manufa turing parameters . . . 76
4.23 High-qualityofpre ursormaterialsobtainedbya ombinationof oldand
hot pressing . . . 76
4.24 Mi rostru ture ofpre ursor materials withoutblowing agent . . . 77
4.25 Pre ursor materialsmanufa tured . . . 78
4.26 Mi rostru tureof pre ursormaterialspreparedof dierent Al-alloy
(mag-ni ationof x100) ontaining0.6wt.%.of TiH
2
. . . 79 4.27 Mi rostru ture of pre ursor materials prepared of Al-A alloy ontaining0.6wt.%.of treated and as-re eived TiH
2
powders . . . 80 4.28 Foaming behaviourofpre ursor material withdierent Al-alloysontain-ing0.6wt.% TiH-1as-re eivedpowder,usinga pre-heatedfurna eat 750
◦
C 81
4.29 Foaming behaviour of Al-A alloy pre ursor material ontaining 0.6 wt.%
TiH-1, using apre-heated furna eat dierent foaming temperatures . . . 82
4.30 Foaming behaviourofpre ursor material withdierent Al-alloys
ontain-ing0.6wt.% TiH-1 treated powder,using apre-heated furna eat 750
◦
C . 82
4.31 Foaming behaviourofpre ursor material withdierent Al-alloys
ontain-ing0.6wt.% TiH
2
treated powder, usinga pre-heatedfurna eat 750◦
4.33 Foaming behaviour of Al-A pre ursor material ontaining 0.6 wt.%
TiH-1Ttreated powder orTiH-2T treatedpowder,usingapre-heated furna e
at 750
◦
C . . . 83
4.34 Pre ursor materials. . . 84
4.35 Mi rostru ture of the dierent pre ursormaterials. TiH
2
parti les (white olour)is distributedinto theAl-A alloymatrix (light gray olour) . . . . 854.36 Al-alloyfoams preparedto me hani al hara terisation . . . 85
4.37 Stress-strainbehaviourasobtained inuniaxial ompressiontestsfor both
AlSi1Mg and AlSi7foams ( ylindri alsamples
∅
=30mm x30mm) . . . 88 4.38 Stress-strain urve inthe linearelasti region . . . 894.39 Averageplateau stress forboth studied Al-alloyfoams . . . 90
4.40 Energy absorption apabilities ofAl-alloyfoams . . . 92
A
Se tion area [m
2
℄E
Young modulus [N.m
2
℄E
∗
Metalli foamsYoung modulus [
N.m
2
℄
E
s
Base materialYoung modulus [N.m
2
℄
E
v
Absorbedenergy pervolume unit [M J/m
3
℄
F
Compression for eapplied [N
℄K
Thermal ondu tivity [W.m
−1
.K
−1
℄K
s
Thermal ondu tivity [W.m
−1
.K
−1
℄P
1
Referen e pressure [P a
℄P
2
Realpressure inthesample [P a
℄R
∗
Foamresistivity [
Ω.m
℄R
s
Base materialresistivity [Ω.m
℄T
Temperature [K
℄V
c
Constant for the alibrationof equipmentV
p
Truevolume of thesample [m
3
℄
V
r
Constant for the alibrationof equipmentη
ef
Energy absorptione ien y [%
℄φ
Contribution of theedgesof the ells onstantρ
∗
Foamdensity [Kg/m
3
℄ρ
r
Relative densityρ
s
Densityof foambasemetal [Kg/m
3
℄
σ
Compression stress [P a
℄ε
Compressive strain [%
℄q
Heattransfer gradient [W
℄Al
AluminiumC
CarbonCa
Cal iumCu
CopperF e
IronH
HydrogenM g
MagnesiumM n
ManganeseN
NitrogenN i
Ni kelO
2
OxygenSi
Sili onT i
TitaniumAl
2
O
3
AluminaSiC
Sili onCarbideT iH
x
(0 < x < 2)
Titanium HydrideT iN
0
.3
OsborniteT iO
2
RutileIntrodu tion
Metalli foamsare urrentlybeing onsideredforanumberofappli ationsinthemarine,
aerospa e and automotive industries. Re ently a number of metal foams have been
de-velopedtorepla epolymerfoamsinappli ationswheremulti-fun tionalityisimportant.
Closed Al-alloyfoams oera unique ombination ofproperties su haslowdensity,high
stiness, strength and superior energy absorption. In addition, this material oers
im-proved soundproong hara teristi s, low thermal ondu tivity hara teristi s, and low
toxi ityunderre onditions. Al-alloyfoamhasbeenusedin ompositeintegral armour
in order to develop a ballisti material. Here, it was found that the foam exhibited
signi ant non-linear deformation under loading and stress wave attenuation [1 ℄.In
ad-dition, the Al-alloy foam armours also resulted in a redu ed dynami dee tion in the
ba king plates. At present, there is an in reasing interest in the potential oered by
metalli foams foruseinhighperforman esandwi hstru tures. Unfortunately,inmany
ases, the skins are adhesively bonded to the metalli ore, a pro edure that in reases
the length of the manufa turing y le and its asso iated osts. Many manufa turing
pro esses, reported in the literature review are still being improved, as it is the ase
of powder metallurgi al method. The urrent resear h studies the onne tions between
te hnologi al fa torsand the propertiesof the obtained foams inorder to improve their
quality. Inorderto obtain ahighqualitymaterial withhighte hnologi al performan es
it isne essarytheunderstanding, orrelation and ontrol ofte hnologi al parameters.
The aimof thisthesiswastodeveloppre ursormaterial withhighqualityto obtain
Al-alloyfoams,using ommer ialpowdersofAl-alloyandtitaniumhydride. Inaddition,the
main me hanisms responsible for the formation of foam, as the thermal de omposition
and oxidation behaviour of the titanium hydride and thethermal behaviour of the
Al-alloywere investigated indetail.
This dissertation isorganised in hapters inorder to fa ilitateits understanding. Thus,
Chapter 2 gives an overview of the urrent state of the art with regard to the
manu-fa turing pro esses, properties and appli ations. Experimental pro edures used in the
present work for pre ursor material produ tion, foam produ tion and the used
experi-mental methods aredes ribedinChapter3. Resultsanddis ussionaregiveninChapter
State of the Art
2.1 Introdu tion
Honey ombs, foams,wood, ro k, plant stems, bone and tissue engineering s aolds all
havea ellularstru turethatgivesrisetouniquepropertiesthatareexploitedin
engineer-ing and inmedi ine. Nature, too, uses ellular materials to provide stru tural support
as well as to ondu t uids. Cellular materials onstitute attra tive lass of materials
with a wide variety of stru tural and fun tional appli ations. These materials an be
madefrommetal, erami andpolymer. Inadditiontotheirlowspe i weight,metalli
foams feature a series of me hani al, thermal and a ousti properties that make them
parti ularly well suited appli ations inthe automotive, biome hani al and onstru tion
industries [2,3℄.
Metalli foamsbelongof this lass of ellular materialsthatareused inengineering and
medi ine appli ations. Its advantages are that these materials are lighter than
tradi-tionalsolidforms(solidsheets),morefun tionalthan traditional ellular materials
(hon-ey ombs, polymer foams), andtherefore anbe more ost-ee tive andenvironmentally
friendly than any ompetitormaterial[4 ℄.
Closed- ell metalli foamwas rst reportedin1926 by Mellerina Fren h patent where
foaming of light metals eitherbyinert gasinje tion or byblowing agent was suggested.
The next twopatents on sponge-like metalwere issuedto Benjamin Sosnikin1948 and
1951 who applied mer ury vapour to blow liquid Aluminium [5 ℄. Closed- ell metalli
foams have been developed sin e about 1956 by John C. Elliott at Bjorksten Resear h
Laboratories [6℄. Although the rst prototypes were available in the 50s, ommer ial
produ tion was started only in the 90s by Shinko Wire ompany in Japan. Metalli
foams are ommonly made byinje ting a inert gas or mixing a blowing agent (usually,
Titanium Hydride)into moltenmetal [7 ℄.
Nowadays,there areanumberof su essfulmanufa turingpro essesfor metalli foams.
Someofthesepro esseshasbeen ommer ialisedinEurope,AsiaandNorth-Ameri aby
ompanies likeAlantum (KoreaandGermany), Cymat(inCanada),DunlopEquipment
(inEngland), Glei hGmbH (inGermany), Mepura(inAustria),Shinkowire(inJapan)
and Re emat(inNetherland) [813℄. Dierent typesof metalli foams are ommer ially
availablesu has,Ni kel,Copper,Zin ,Steel,AluminiumandGold. Amongthemetalli
highdu tility, highthermal ondu tivity,and ompetitive ostof themetal[2℄.
The most important onferen e in this eld is the international onferen e on Porous
MetalsandMetalli Foams seriesthatisheldregularlyattwo-yearsintervalssin e1997.
The onferen es have been s heduled to provide a forum for resear hers a tive in the
elds of porous and foam materials and for industrial materials engineers and produ t
designers seeking newmaterials [7 ℄.
2.2 Classes of metalli foams
Metalli foams belong to a group of materials alled ellular solids. Cellular solids are
denedashavingaporosity>0.7[3℄. Naturalfoamsareprodu edbyplantsandanimals
su h as ork orbone. Manmadefoams an be manufa turedfromavarietyofmaterials
su has erami s,polymersandmetals. Cellularmetalli materialsarebeingin reasingly
regarded as a solution for problems of light weight onstru tion, passive safety, sound
damping or ltering purposes. Foams are three-dimensional arrays of ells that an be
dividedinto two ategories: open- elland losed- ell metalli foams[3℄,asshowninFig.
2.1. Intheopen- ellfoams,the ellsareinter onne tedby elledges,orligaments. Inthe
losed- ellfoams,the ellsarepredominantlyisolatedfromea hotherbysolidfa es. This
hara teristi an dire tlybeobserved by opti almi ros ope or an also be determined
from the permeability of foam to a gas or a liquid. Open- ell metalli foams allow
the passage of liquids and gases for dierent appli ations ranging from ltering to heat
ex hange andgive thefoamits in reasedsurfa e areawhile the losed- ell onguration
isoptimal forenergy absorptionand stru turalappli ations likein arbumpers, bridges
and buildings. Open- ell metalli foams are usually used to enhan e heat transfer in
appli ationssu has ryogeni heatex hangers,heatex hangers forairborneequipment,
oal ombustors, ompa t heatsinksfor highpowerele troni devi es,heatshielding for
air raftexhaust, ompa theatex hangers,liquidheatex hangers,air- ooled
ondenser- ooling towers and regenerators for thermal engines[2 , 4 , 7 , 14 , 15℄. A uid media an
not pass through losed- ell foam. Closed- ell foams are being used in the transport
industry,aslightweight stru tures,energy-absorptionstru turesanddamping stru tures
[7℄. In thiswork the foams ofinterest are losed- ell Aluminiumalloyfoams.
(a) (b)
Figure 2.1: Metalli foams. a) open- ell stru ture (NICOFOAM Ni kel foam) [16℄. b)
2.3 Manufa turing Pro esses
Nowadays,therearemanywaystoprodu emetalli foamsorsimilarporousmetal
stru -tures. Themanufa turingpro esses anbe lassieda ordingtothestateofthestarting
metal: liquid, powdered and ionised - asshown in Fig. 2.2 [3, 4 , 7, 14, 15, 1820 ℄.The
most important advantages and disadvantages for the most ommon methods of these
materials arepresented.
Figure2.2: Manufa turingpro esses ofmetalfoams [3, 4,7℄.
2.3.1 Produ tion of foams from metalli melts
Metalli melts anbedire tlyfoamedbytheinje tinggasesorbyaddingablowingagent
thatreleasesgasthroughitsthermalde omposition. Otherindire tmethods an alsobe
usedthroughapolymeri foamorby astingtheliquidmetalaroundsolidllermaterials
whi h reserve spa e forthe pores or whi h remaininthefoam[3, 4,7℄.
2.3.1.1 Dire t foaming methods
There aretwoways fordire tlyfoaming metalli melts whi harealreadyinthestageof
a large-s ale ommer ial exploitation. The rst one is exploited by Cymat Aluminium
Corporation (in Canada) in whi h metalli melt is foamed dire tly by inje ting gases
into themoltenmetal(Fig. 2.3). Thegases(e.g. Air,Nitrogen,Argon)areinje ted into
the molten metal using spe ially designed rotating impellers. These impellers have to
produ e very ne gas bubbles in the molten melt and distribute them homogeneously.
The gasbubbleswhi h arethenformed into themolten metaltendto riseto its surfa e
qui kly due to the high buoyan y for es in the high-density liquid but this rise an be
impeded by in reasing the vis osity of the molten metal. This an be done by adding
ne erami powders(e.g. Alumina, Sili on Carbide and Zir onia) or alloying elements
whi h form parti les in the melt. The formed foamy mass is ontinuously drawn o
by means a onveyor. An additional upper onveyor is used to reate a at surfa e
foams, ontinuously. A possible disadvantage is the eventual ne essity for utting the
foam due the high ontent of erami parti les(10-30 vol. %)used inthepro ess. The
main disadvantage of this pro ess is the poor quality of the foams produ ed. The ell
size is large and oftenirregular, and thefoams tendto have a marked densitygradient.
Foam panels with 1m in width and thi kness range of 25-150 mm, an be produ ed in
ontinuous length at produ tion rates of 900 kg/hour. The relative densities range of
these foams is0.05-0.55 g/ m
3
. Theaverage ell size is2.5-30 mm [3,7℄.
Figure2.3: Dire tfoaming of meltsby gasinje tion[3℄.
The se ond alternative method is exploited by Shinko-Wire Company (in Japan) using
Alporasas the trade name(Fig. 2.4). Thismanufa turing pro ess, themetalli meltis
foamed by adding blowing agent parti les. The blowing agent parti les(e.g. Titanium
Hydride) de omposeundertheinuen eof heatand releasesgaswhi h thenpropels the
foaming pro ess. Cal ium is previously added to the molten metal for adjust the melt
vis osity. After the vis osity hasrea hed thedesiredvalue, TitaniumHydride (TiH
2
)is added(typi ally1.6wt.%),asablowingagentbyreleasingHydrogen(H2
)gasinthehot vis ous liquid. The melt starts to expand slowly and gradually lls thefoaming vessel.After ooling the vesselbelow the melting point of thealloy, theliquid foam turns into
solid Al foam. Foam blo ks is is removed from the mould and are are ut into sheets
of the required thi kness. This pro ess is apable of produ ing large blo ks of good
quality. Blo ks with450 mm inwide, 2050 mm inlength and 650 mm inheight an be
produ ed. Thesefoams have uniform pore stru ture and do not require the additionof
erami parti les, whi h makesit brittle. However, the method is more expensive than
foaming melts by gas inje tion method requiring more omplex pro essing equipment.
The density rangeofthese foams is0.18-0.24g/ m
3
,and themean ell size isabout4.5
mm.
2.3.1.2 Solid-gas eute ti solidi ation
This method whi h was developed some years ago at theDnepropetrovsk Metallurgi al
(a) (b)
( ) (d)
Figure 2.4: Dire t foaming meltsbyaddingblowing agents[3℄.
withHydrogengas. The metalto be foamedismeltedinan auto lave witha ontrolled
pressure of Hydrogen, sothat themelt be omessaturated with Hydrogen. The melt is
thendire tionallysolidiedandasit oolsthroughthesolid-gaseute ti point,itbe omes
supersaturated. Atwo phasesolid/gas mixture is simultaneously formed fromthemelt,
yielding an anisotropi porous solid with ylindri al pores oriented in the solidi ation
dire tion. Metalli foams manufa tured by this pro ess are alledas GASARs[3℄. The
main advantage of this pro ess is to allow a ontrol of the nal properties of foamed
materials, namelythegeometry,the sizeandtheorientation ofthe ellularpores. Radial
and axial pores an be produ ed using a ylindri al asting vessel. The possibility of
solidifying the liquid dire tionally oersthe advantage of making foams with elongated
pores. However,the maximum porosities whi h an be a hieved by this pro essare not
very high(5-75%) butmetalswithmediumandhighmeltingpointssu hasCopperand
Ni kel an be foamed. Layered stru tures with alternating bands of solid and porous
materials an alsobeprodu ed [3℄.
2.3.1.3 Investment asting
Another method of foaming of metalli melts applies investment- asting te hnique for
open- ell metalli foamprodu tion (Fig. 2.6). Commer iallyknownDUOCEL
T M
foam
usedinheatex hangersismanufa tured bytheinvestment asting. TheERGMaterials
and Aerospa e Corporation hasbeen manufa tured for the aerospa e, national defense,
semi ondu tor manufa turing, biote h andother high te hnology industries. A erami
repli ate is produ ed using a polymeri foam pre ursor and then the liquid metals are
penetratedandrepla ethepolymeri foam. Awideregularshapeanduniformityoffoam
stru ture anbepossible withthis method. Thepro essprodu esrelatively small
quan-tities of expensive, high-quality foams withreliable material properties. The porosities
an be as high as98%, withporesize between one and several millimeters. Any metal
or alloy an be foamed; however, its widespread use is limited due to the omplexity,
ost and di ulty of s aling up the pro ess [3, 4, 7℄. Duo el
R
Aluminium foam is our
most popularmetalli foam, followed by opperfoam. Other Duo el
R
foammetalsthat
we have produ ed, but arenot urrently ommer ial produ ts in lude tin, Zin , Ni kel,
In onel, Silver, and Gold [2℄.
(a) (b)
( ) (d)
Figure 2.6: Investment astingpro ess [3 ℄.
2.3.1.4 Synta ti foams using ller materials
Themost ommonmanufa turingmethodformetalli synta ti foamsismeltinltration,
where themolten metalispressure inltrated into a randompa kof erami spheresor
hollowspheres(Fig. 2.7). Ava uumoranexternalpressure anbe reatedtofa ilitythe
meltinltration. Thevolume per entage of the metalmatrix is thus determined by the
amount ofinter-parti le spa eofthe erami mi rospheres. Ifthe erami sphereshavea
similar parti le size,the metal volume per entage isxedat about 37%. Therefore, the
fabri atedsynta ti foams(fabri atedwithasimilarparti lesizeof erami spheres)have
asimilarvolumefra tionofmetalmatrix. Thevolumeper entageofmetalmatrixofthe
foam anbede reasedbyembedding erami sphereswithmultimodalsizedistributions,
in ludingAluminium,Zin ,Magnesium,Lead,Silveret . [7℄. Partsofapredenedshape
an be produ ed usingamould oftheappropriate geometry. The erami spheres ould
beremoved by lea hingthem insuitable solvents or a idsor thermal treatment [4℄.
Figure2.7: Syntati foams usingller metal[4℄.
2.3.2 Foams made from metal powders
Metalli foams anbealsoprodu ed usingmetalpowders. Someofthesemanufa turing
pro esses arealsointhestate of ommer ial exploitation,but insmall-s ale.
2.3.2.1 Powder metallurgy
The Powder Metallurgi al method (PM method) is one of the ommer ially exploited
methods toprodu e losed- ellmetalli foams. Thismanufa ture pro esswasdeveloped
and patentedat FraunhoferInstituteinBremen(Germany)[3,4℄. ThisPMmethod an
beusedto produ e foams of dierent metalsand itsalloys [21℄,su h asAluminiumand
its alloys, Tin, Zin , Lead, Steel and Gold whi h is one of the advantage of this PM
method.
Metal powders (elementary metal powders, alloy powders or metal powder blends) are
mixedwithasmallfra tionofasuitablepowderedblowing agent by onventional mixers
(e.g. turbula mixer). The blowing agents usually used for produ ing Al-alloy foams
using thePMmethod aremetalhydrides, su hasTitaniumHydride (TiH
2
),Zir onium Hydride (ZrH2
) and Magnesium Hydride (MgH2
). The employment of others blowing agent powders, su h as the arbonates, as a ost-ee tive alternative to metal hydridesblowing agent hasbeen also investigated. If metalhydrides are usedas blowing agents,
a ontent oflessthan 1%is su ient inmost ases [7,22 , 23℄.
In thissimple manner,a homogeneous powdermixture isobtained. Subsequent to
mix-ing,thepowderblendishot ompa tedby onventionalpressingte hniques(e.g. uniaxial
pressing,isostati pressing,extrusion)toa dense,semi-nishedprodu t alledfoamable
pre ursor material or pre ursor material in whi h the blowing agent parti les must
be homogeneously distributed into the metalli matrix. The temperature of this
om-pa tion isneartothe initial ofthermalde omposition temperature oftheblowing agent
material. This foamable pre ursor material an be pro essed into sheets, rods, proles
et . by onventionalte hniqueslike rollingor extrusion.
Figure2.8: Powder metallurgy
pro ess [24 ℄. Finally, this foamable pre ursor material is heated to
temperatures above the melting point of the matrix
metal, resulting in thefoam itself. The metal expands
developing ahighlyinternal porous stru tureof
losed- ells dueto the simultaneouso urren eofthemelting
of the metal and thermal de omposition of the
blow-ing agent withtherelease of a gas(e.g. H
2
inthe ase of metal hydrides). The liquid foam is then ooled inAir,resultinginasolidfoamwith losed ellsandwitha
verythindenseexternalskinthatimprovesthe
me han-i al properties of thesematerials. Thefoaming pro ess
usually takespla e into the losed mouldswith asame
design and dimensionsof thenalfoampart[2, 3℄.
The densityof metalli foams an be ontrolled by
ad-justing the ontent of blowing agent and several other
foaming parameters. The high quality foams of these
dierent metals an be obtained by hoosing the
ap-propriate blowing agent. Moreover, the manufa turing
parametersofmixture,the ompa tionandthefoaming
y le, all have to be hosen arefully [7, 22 , 23℄. This
pro ess anprodu efoamswithporositiesbetween75%
and 90%.
Themain advantageofthis PMmethod isthe
possibil-ity to produ e omponents of metalli foams with
dif-ferent ar hite tures (e.g. sandwi h systems, lled
pro-les and 3D omplex shaped stru tures) in
ompari-son to theothers [25 , 26 ℄. Thematerials an be mixed
during thefoaming stepwithoutjoining adhesives[22 ℄.
Other advantage lies in the fa t that the addition of
erami parti les are not required, avoiding the
brit-tleme hani albehaviourthatthis parti lesinfer tothe
foams. Moreover,the foampartsare overed by an
ex-ternal dense metal skin that improves its me hani al
behaviour, providing a good surfa e nish. The disadvantage of this PM pro ess is the
high ee tive ost of thepro ess whi h ismainly asso iated to the ost ofthe powders.
Another disadvantage isthe di ulty manufa ture large volume foam parts.
Neverthe-less,sandwi hpanelof2mx1mx1 m an alreadybemanufa tured. Furthermore,itmust
be pointed out that during PM method it is still rather di ult to fully ontrol the
foaming pro ess, whi h results in la k of uniformity of the pore stru ture. The foams
obtained by this manufa turing pro ess present a losed- ell stru ture and an external
2.3.2.2 Foaming of slurries
Metals anbefoamedbypreparingaslurryofmetalpowdermixedwithablowingagent.
Theslurryispouredinto amouldaftermixinganddriedthereatelevatedtemperatures.
Theslurrybe omesmorevis ousand startstofoamasgasbeginstoevolve. Ifsu ient
stabilising measures have been taken the expanded slurry an be dried ompletely thus
obtaining a metalli foam. Su h foams have been produ ed from aluminium powders
using orthophosphori a idwithaluminiumhydroxide or hydro hlori a idasa blowing
agent. Relative densities down to 7% have been a hieved but there are problems with
low strengthand ra ksinthe foamed material [7, 20 ℄.
2.3.2.3 Gas entrapment
Metals anbefoamedwithoutusingablowingagentby ompressingpowderstoa
pre ur-sormaterialand allowing gasto beentrappedinthemetalstru ture during ompa tion
(Fig. 2.9). Heating the pre ursor materialthen leads to an expansion of themetal due
to the internal pressure reated bytheentrapped gas. Thepro ess hasmainly been
de-signedformakingporoustitaniumstru turesforsomeair raftappli ationsmanufa turer
Boeing(USA).Forthistitaniumpowderislledintoa ontainerwhi histheneva uated
andrelledwithargongas. Thelled ontaineristhendensiedbyhotisostati pressing,
subsequently worked and nallyfoamed bymeans of anappropriate heattreatment.
Figure 2.9: Gas entrapment te hnique. [3℄.
2.3.2.4 Other Te hniques
There are many ways to make porous metalli produ ts from metal powders, bers, or
hollowspheres. Theeasiest method istosinterloosepowderllings ina anister togive
a porous material withopen porosity. Mixtures of metal powders and polymer binders
an be extruded and then heat treated to produ e porous parts. In this way porous
materials with ylindri al pores an be produ ed. Rea tion sintering of metal powder
mixtures analsoyieldporousprodu ts. Hollowspheresmadeoftitaniumorsteel anbe
usedtoformhighlyporousstru turesbysintering. Orderedanddisorderedarrangements
still su ient for many appli ations. By inltrating the intersti es between the hollow
spheres, the strength an bein reased [3 ℄.
2.3.3 Produ tion by Deposition Te hniques
Deposition te hniques start from the ioni state of metals. The metal is galvani ally
deposited on a polymer foam withopen ells. Thispro ess and the investment asting
pro ess therefore have in ommon that the a tual foaming does not take pla e in the
metalli statebutwithapolymerwhi histhenrepla edbyametal. Galvani deposition
onapolymerfoamrequiressomeele tri al ondu tivityoftheinitialpolymerfoam. This
is a hieved by dipping the polymer foam into graphite solutions or by oating it with
a thin ondu tive layer bymetal vaporisation. After ele troplating thepolymer an be
removed from the metal/polymer omposite by thermal treatment. Foams of various
grades an be fabri atedranging from 2 to 30 ellsper m(6 to 70 ppi). Thepreferred
metal isni kel or ani kel- hrome alloybut opperfoams an also be fabri ated. Foams
have been oered ona ommer ialbasisunderthenameRETIMET (Dunlop Ltd., GB)
and CELMET (Sumitomo, Japan). This method isinvolveshighprodu tion osts[2℄.
2.4 Properties
Metalli foams ombine properties of ellular materials with those of metals. For this
reason, metalli foams are advantageous for lightweight onstru tions due to their high
strength-to-weight ratio, in ombination with stru tural and fun tional properties like,
rash energy absorption, sound and heat management. Fig. 2.10 shows an overview of
properties ofthe metalsfoams andits positioninrelation withothermaterials. In
om-parison to polymer foams (foruses inautomobiles), metalli foams arestier, stronger,
and more energy absorbent. They are more re resistant, and have better weathering
properties when onsidering UV light, humidity, and temperature. However, they are
heavier, more expensive,and non-insulating.
Manymetalsandtheiralloys anbefoamed. Amongthemetalli foams,Al-alloysfoams
are ommer ially the most exploited ones due to their low density, high du tility, high
thermal ondu tivity, and ompetitive ost ofthe metal. The ranges ofmain properties
oered by urrently availablemetalli foamsarelisted inTable 2.1:
Relative density, foam morphology and poresize depend on the manufa turing pro ess
usedtoobtainametalli foam,asshownisFig. 2.11. These hara teristi sae tphysi al
Figure2.10: Propertiesof ellular anddense engineering materials[2 ℄.
Figure 2.11: The range of ell size and relative density for the dierent metalli foam
2.4.1 Density
Therelativedensityisthemostimportant hara teristi whi hae tsme hani al
proper-tiesoffoams,su hasYoungmodulus,stressplateauandenergyabsorption. Therelative
density (
ρ
r
) is the ratio of foam density(ρ
∗
) to solid material density (
ρ
s
) and is given byρ
r
= ρ
∗
/ρ
s
(2.1)andrelates to porositysimplyby:
P orosity = 1 − (ρ
∗
/ρ
s
)
(2.2)Thereisnodoubtthatboththedensityofametalli foamandthematrixalloyproperties
inuen e, e.g. modulus and strength of the foam. Figure 2.12 shows the ee t of the
densityonthe Youngmodulus. Densityvariationandimperfe tionsyieldalarges atter
ofmeasuredproperties,whi hisdetrimentalforthemetalli foamsreliability. Me hani al
studiesdemonstratethatsele tivedeformationoftheweakestregionofthefoamstru ture
leads to rush-band formation.
Figure 2.12: Young modulus for dierent foam.
2.4.2 Stru tural properties
Thereareseveralstru turalparametersfor hara terisingthemetalli foams(Fig. 2.14),
namely number, size-pore distribution, average size, shape and geometry of the pores,
of theexternal dense surfa e for des ribing the ellular ar hite ture of the foams. The
propertiesofthemetalli foamsareinuen edbythesemorphologi alfeatures. Fig. 2.15
show some defe ts or imperfe tions of a losed- ell foam that de rease the me hani al
properties [2,3℄.
Figure2.13: Listof parameters to des ribe the stru ture ofmetalli foam[21℄.
Figure 2.14: Closed- ell Aluminiumfoam[21 ℄.
Progress hasbeen madeinunderstanding the relationshipbetween properties and
mor-phology. Although this exa t interrelationship isnot yet su iently known, one usually
assumes that the properties are improved when all the individual ells of a foam have
similar size and a spheri al shape. This has not really been veried experimental. All
studies indi ate that the real properties are inferior than theoreti ally expe ted due to
stru tural defe ts (Fig. 2.14). This demands a better pore ontrol and redu tion in
Curved,wiggledormissing ell-walls(seeFig. 2.14)mi roporesonthe ell edgesand ell
wallsand nonuniformities densityarethefurther imperfe tions degrading thestrength,
and inturn, result in a redu ed deformation energy absorbed under ompression. Cell
morphology and inter onne tion ould also ae t thermal and a ousti properties. It
is widely a epted that foams with a uniform pores-distribution and defe ts free are
desirable. This wouldmake theproperties more predi table. Onlythen, metalli foams
willbe onsideredreliablematerialsforengineeringpurposesandwillbeableto ompete
with lassi al materials. Despite their quality improvement in the last 10 years the
resultingmetalli foamsstillsuerfromnon-uniformities. The ellsizeislargeandoften
irregular, and the foams tendto have a marked density gradient (Fig. 2.15). S ientists
aim to produ e more regular stru tures withfewer defe ts in a more reprodu ible way,
whi h will be oneof the motivations of the urrent resear hinthis eld[2, 3℄.
Figure2.15: Closed- ellfoamwithgradient density[21℄.
2.4.3 Me hani al Properties
The me hani al properties of the metalli foams have been evaluated using both stati
anddynami ompressiontests. Asimplewaytoobtainmaterialdeformationbehaviour
under an imposed load is to perform a stati test. The most ommon types of testing
apparatuses employhydrauli , pneumati or servo-me hani al powerto ompress or
ex-tend the spe imen. Typi ally, to be onsidered a stati test, the strain rate must be
less than 10
−1
s
−1
. The engineering strain rate is a fun tion of rosshead speed and
spe imenheight. Theimportantproperties obtainedfromthestati testsaretheelasti
modulus, yieldstrength, plateaustrength andthe absorbed energy ata ertainstressor
strain. Dynami testing is performed to determine the behaviour of materials at high
strain rates. This knowledge is important when designing for rash or blast loadings.
The deformation under ompressive loads, elasti deformation, ollapse, plateau stress
and energy absorptionare amongwidelystudied properties offoams [3℄.
2.4.3.1 Stress-strain behaviour under ompression
Foamsshowa hara teristi stress-strainbehaviorunder ompressiveloads. Compressive
densi ation,asillustratedinFig. 2.16. Theelasti modulus, theplateaustressandthe
densi ationstrainarethemost important me hani alparameters whi haredetermined
using these urves. These me hani al parameters in rease with in reasing the foam
density, while the defe ts or imperfe tions have detrimental ee ts on the me hani al
properties. The ompressive stress behavior in reases with in reasing the foam density
(Fig. 2.17).
Figure 2.16: Compressive stress-strain urvesof a) elastomeri , b) elasti -plasti and )
elasti -brittle foam.
Figure2.17: Ee t of foamdensityon ompressivestress-strain urves[14 ℄.
Linear elasti region Therstregion ofthestress-strain urve islinearelasti region
in whi h the stress in rease almost linearly with the strain. The deformation in this
region is ontrolled by ell wallbendingand/orstret hingdependingonthestru tureof
the foams: open or losed ell foam. Open- ell foam of low relative densities (the ratio
between foamdensityand solidfoammaterial density(
ρ
∗
wall bending. Within reasing relative density(
ρ
∗
/
ρ
s
>0.1), ell edge ompression plays a signi ant role. Fluidowthrough open- ell foam ontributes to theelasti moduleifthe uid has a high vis osity or thestrain rate is ex eptionally high. Besides ell edge
deformation, the thin membranes of the losed ell foams, whi h form the ell fa es,
stret h normalto the ompression axisand therefore ontribute to the modulus. If the
membranes do not rupture, the ompression of the ell uid trapped within the ells
also in reases the modulus. Ea h of these me hanisms ontributing to the linear-elasti
responseofthe foamsis showns hemati allyinFig. 2.18for open and losed- ell foams
[2, 4,7℄.
(a) (b)
Figure2.18: Deformationme hanisms offoams: a) open- ell foam,sequentially ell wall
bending, ellwallaxialdeformationanduidowbetween ellsandb) losed- ellfoams,
sequentially ell wall bending and ontra tion, membrane stret hing and en losed gas
pressure [7℄.
GibsonandAsbhy(1997)developasimple ubi unit ellmodeltopredi tthenormalized
me hani al properties of the foams, as shown in Fig. 2.19 However, the stru ture and
shapeofthe ellsaremore omplexthanthoseofthe ubi model. Thedeformationand
failure me hanisms of the ubi model are however quite similar to those of real foams
and thereforeit isvery usefulinpredi ting me hani al properties.
(a)Cubi unit ellmodelsofopen- ell. (b)Cubi unit ellmodelsof losed- ell.
Figure 2.19: Cubi unit ell models of open- ell and losed- ell foams developed by
The elasti modulusis an important parameterto determine inthis region whi h is the
slopeofthis urveregion. There aremanymodelstopredi tthiselasti modulus. Based
on thissimple ubi unit ell model,Gibson andAsbhyproposedthefollowing equation
of the elasti modulus of the open ell foams (E
∗
), whi h is al ulated from the
linear-elasti dee tion of a beam of length l loaded at its mid point by a load F, is given
as,
E
∗
E
s
= C
1
ρ
∗
ρ
s
(2.3)where s refers to thesolid material from whi h thefoam is made and C
1
is a onstant. The experimental elasti modulus of open- ell foams showed that this onstant (C1
) is nearlyequaltounity. GibsonandAsbhyproposedthefollowingequationforthemodulusof imperfe t losed- ell foams
E
∗
E
s
≈ φ
2
ρ
∗
ρ
s
2
+ (1 − φ)
ρ
∗
ρ
s
(2.4)where
φ
is the fra tion ofthe material ontaining on ell edges. The elasti modulus of the losed- ell foams based on this ubi model that in ludes en losed gas pressure isgiven by
E
∗
E
s
= C
1
φ
2
ρ
∗
ρ
s
+ C
1
′
(1 − φ)
ρ
∗
ρ
s
+
P
0
(1 − 2ν
∗
)
E
s
1 −
ρ
ρ
∗
s
(2.5)where
φ
isthe fra tionof the solidwhi h ontained inthe ell edges having a thi kness (te) andtheremainingfra tion(1-φ
) inthe ell fa esofathi kness(tf),P0
istheinitial pressure of the ell uid and C1
and C1
′
are the onstants. The rst, se ond and third terms of these equation are the ontribution of ell wall bending, membrane stret hingand en losedgaspressure, respe tively.
TheYoungmodulusdependsaboveallonthefoamdensityinwhi htheirvaluesin rease
withthein rease ofthedensity,asshowninFig. 2.20.
Plateau region Collapse regionpro eeds withastress plateau either witha onstant
value or in reasing slightly with strain. Linear elasti ity is generally limited to small
strains. Elastomeri foams an be ompressed mu h larger strains. Deformation is still
re overable, but non-linear. In ompression, the stress-strain urve shows an extensive
plateau at the elasti ollapse stress (
σ
∗
el), see Figure 2.16 (a). Foams made from
materialthathaveaplasti yieldpointsu hasrigidpolymersanddu tilemetals ollapse
plasti allywhenloadedbeyondthelinear-elasti region. Severaldeformationme hanisms
o ur is this region su h as elasti bu kling and brittle rushing of the ell walls and
formationof plasti hinges. Plasti ollapsegivesalonghorizontalplateauinthe
stress-strain urvesimilar totheelasti bu kling, but thestrainisno longer re overable. Both
elasti bu kling and plasti failure are lo alized; a deformation band is usually formed
transverse to the loading axis and propagatesthrough undeformed se tions of thefoam
within reasing strain until all thefoamse tionis lledwiththeband.
Figure2.20: Variationof Young'smodule withdensity[3℄.
region where itinitiates the plasti regime. Tryingto estimate thevalue of theplateau
tension, an equation was introdu ed by Gibson and Ashby. This equation takes into
a ount the study of a foam with a ubi geometry and with a losed stru ture. This
stressdependsonthedensityofthefoamandtheyieldstrength(
σ
ys
)ofthesolid[3,7,14℄.σ
pl
σ
ys
≈ 0.3
φ
ρ
∗
ρ
s
2
3
+ 0.4 (1 − φ)
ρ
∗
ρ
s
(2.6)Where
σ
pl
isthe plateau stress.Theplateau stressalsoin reases withthefoamdensityin rease(see Fig. 2.17). The
de-terminationoftheplateaustressvalueisimportantto hara terizetheenergyabsorption
apa ityof foams [21℄.
Densi ation region Following theplateau region, at a riti al strain, the ell walls
start to tou h ea h other and the densi ation region begins. The stress inthis region
in reases rapidly and approa hes to the strength of thesolidfoam material. Foam
den-si ation o urs atthe densi ation strain(
ε
D
). Thepointat whi hthefoamstartsthe densi ation isnot well dened and the denition isfound to varybetween resear hers.Commonmethods ofdeterminingthedensi ation strain(orstress)basedonthe
stress-strainresponsearebyvisualinspe tion,atastress1.5timesthestressvalueatastrainof
0.5. Other methodutilizesthepointofinterse tionoftheslopeoftheplateauregionand
thatofthedensi ation regionasthepointofdensi ation. Other resear hers hoosean
arbitrarystrainand utilizethis asthedensi ation point. Instru tural appli ations the
densi ation point isnot asimportant asthe yieldpoint. Howeverinenergy absorption
appli ationsthepointofdensi ationisimportantand ompressionbeyonddensi ation
is to be avoided due to the sharp in rease in stress. The densi ation strain (
ε
D
) also in reases within reasing of the relative density,asillustrated inFig. 2.17).2.4.3.2 Energy Absorption under ompression
The possibility of ontrolling the stress-strain behaviour by an appropriate sele tion of
the matrix material, ellular stru tures and relative density makes metalli foams an
ideal material for energy absorbing stru tures. The quality of the energy absorbers is
dened by the ability to absorb energy without themaximum or the highest o urring
a eleration ex eeding theupperlimiting atdamages o ur.
The longstress plateau typi al ofthe stress-strain urves ofmetalli foams gives rise to
ex ellentenergy absorptionproperties[2℄. Damage toanobje tis ausedwhena riti al
for e(ora eleration) levelisex eeded[2℄. Theabilitytoabsorbenergy atafor ebelow
this riti allevelisparamounttotheprote tionoftheobje t[2℄. Thefor eofanimpa t
is dire tly related through geometry to the stress inthe foam [2 ℄.Figure 2.6 shows why
foams aremu h better than solidmaterials at providing damageprote tion. For agiven
stress foams always absorb more energy than a solid due to the bending, bu kling and
fra ture of the foam ell walls[2℄.
Figure 2.21: Stress-strain urvesfor both an elasti solidand afoamshowing the
dier-en einenergy perunitvolume absorbedat the samestress level[2℄.
The ompression of Aluminium foamunder an applied for e results in work. The work
perunit volume (W) up to a strain of
ε
is thearea under the stress-strain urve. It is al ulated inthe following manner:W =
Z
ε
0
σ (ε) dε
(2.7)Little energy absorption o urs intheinitial linear elasti region. Energy absorption in
theplateauregiona ountsforthemajorityofenergyabsorbedbythespe imen. Plasti
deformation ofAluminiumfoamsatnear onstantstress intheplateauregiontranslates
into energy dissipated at near onstant stress. Ideally for better energy absorption the
in rease inenergy absorptionis a ompaniedbylarge stressin reases.
Considering impa t onditions (i.e. strain rates intherange of 10
1
to 104
s−1
), thereisan optimum foam density to absorb energy e iently [2℄. This is illustrated inFigure
2.22 where three foams ofdierent relative densities. Thestress and strainat whi h an
amount ofenergy Wisabsorbed isshown for all threedensities. Thisshows thatiftoo
weak of a materialis hosentherequired amount of energy absorbed ismore than that
under the plateau, the foam densies and the for e in reases sharply before all energy
is absorbed. If however too strong of a material is hosen the load be omes too large
beforeall therequired energy isabsorbed. Themost e ient foamisshownto have the
middledensitywherethefullplateauisemployedinenergyabsorption. Themiddlefoam
absorbs the amount of energy Wat thelowest peak stress.
Figure 2.22: Stress-strain urvesof foams for threedierent density [28℄.
The e ien yof the foamsenergy absorption an be al ulated omparingtheobtained
urve of foam ompression with an ideal urve (Fig. 4.40). This ideal urve has a
re tangular urvethatrepresentsalwaysthesame ompressivestress. Thedetermination
oftheabsorptione ien yisane essarypro edureto hara terize theenergyabsorption
infoams. Energy absorptione ien y (
η
ef
) anbe al ulated bythefollowing equation [7, 21℄:η
ef
=
R
s
′
0
F
s
′
ds
F
max
(s
′
) ds
(2.8)where
F
istheappliedfor e,s'isthedeformationandF
max
isthemaximumfor eapplied abovethedeformations
′
.
The hoi eofenergyabsorbers anbemadethroughtheenergyabsorptiondiagrams(Fig.
2.24). For that,samplesoffoamwitharangeofdensitiesaretested in ompression, ata
xedstrainrateandtemperature. Theareaunderea h urveismeasuredandrepresents
theenergy absorbedperunitvolume(W/V).Thus,thediagram onsisttoplotthevalue
Figure2.23: Comparisonof the ideal absorberand a realabsorber[7 ℄.
standard strainrate and temperature. Ea h foamhas a ompressive stress for whi h it
is the best hoi e. It is assumed that a near ideal foam absorbs a given at a minimum
stress.
Figure2.24: Maximumstresso urringwhenagivendeformation energyisabsorbedby
Aluminium foamswithvarious densities[23℄.
Energy absorption is ae ted by the shape of the stress-strain urve. The in rease in
plateau stresses due to in reased relative densities result in higher values of energy
ab-sorbed. However, the energy absorption e ien y is unae ted by relative density but
hanges with loading dire tion in anisotropi foams. The defe ts in losed ell foams
de rease the energy absorption e ien y. Density gradients within the foam also
de- rease the e ien y ofenergy absorbed. By areful ontrol of foamprodu tion areable
to produ e losed- ell Alporas
R
withsmall ell sizesand,redu eddensitygradients, ell
wall urvature and orrugation. The modied foam with fewer defe ts not only shows
an in reaseinenergy absorption over the unmodiedfoamwithsimilar relativedensity,
2.4.4 Thermal properties
The main thermal properties are the melting point, spe i heat, thermal expansion
oe ient, thermal ondu tivity,surfa e emissivityandthermal sho kresistan e. These
properties in the foams are similar to the base material properties. However, there is
foams thatthese propertiesdepend ontheir ellular stru tures.
Themetalli foams an beasso iated toalowthermal ondu tivitydue totheir ellular
stru tureparameters. Theporesize andthevolume fra tionleadtoasuppressionofthe
onve tion and radiation ee ts, whi h makes metalli foams an interesting material to
use in appli ations on erning thermal insulation [7 ℄. Thermal insulation is the
redu -tion of the ee ts of the various pro esses of heat transfer between obje ts in thermal
onta t or in range of radiative inuen e. The thermal ondu tivity of metalli foams
is dire tly related to the outer surfa e of the foam and is, thus, also dire tly related to
the ondu tivityofthe basematerial. Thevalueof thermal ondu tibilityisalsorelated
to the density value, and thevaluein reases within reasing foamdensity. Thethermal
ondu tivity (K) isgivenbytheFourrier law[29℄:
q = −K.A
dT
dx
(2.9)Where (q) is the rate of heat transfer, (
dT
dx
) is the heat transfers gradient and A is the se tion area of the heat ondu tion. As referred above, thermal ondu tivity (K) analsobegivenapproximatelybyanequationrelatingwithfoamdensityandbasematerial
density[3℄:
K = K
s
ρ
∗
ρ
s
r
(2.10)Where K
s
is thermal ondu tivityof thebase metal, K∗
is the thermal ondu tivity of
metalli foams and r is the value between 1.65 and 1.8. Figure 2.25 shows thethermal
ondu tivity ofavailablemetalli foams.
2.4.5 A ousti properties
The ombinationof good a ousti and me hani al propertiesmakesmetalli foams
par-ti ularly attra tive produ ts. Three ases anbedistinguished withrespe tto thenoise
attenuation materials:
(i) materials for soundinsulation: anymeans of redu ing theintensityof sound. The
performan e of sound insulation material is des ribed in terms of sound
redu -tion index whi h isproportional to there ipro al value of thesoundtransmission
oe ient inlogarithmi s ale.
(ii) materials for sound absorption: Absorbing sound spontaneously onverts part of
the sound energy to a very small amount of heat in the intervening obje t (the
absorbing material), rather than sound being transmitted or ree ted. There are
severalwaysinwhi ha material an absorbsound. The hoi eof soundabsorbing
material will be determined by thefrequen y distribution of noise to beabsorbed
Figure 2.25: Thermal ondu tivity plottedagainstvolumetri spe i heatfor urrently
availablemetalli foams[3 ℄.
(iii) materials for damping an redu e thea ousti resonan e intheair, or me hani al
resonan einthestru ture of theroomitselfor things intheroom.
The a ousti absorption properties ofmetalli foams depend mainly on foamproperties
(parameters)su hasporosity,poremorphology,poresizeandair-owresistan e.
Closed- ellmetalli foamsarepoorsoundabsorbersbutbothopenmetalli foamsaregoodsound
absorbers and they are urrently used for noise redu tion. Open- ell foams are highly
ee tivenoiseabsorbersa rossabroadrangeofmedium-highfrequen ies. Performan eis
lowerat lowfrequen ies. Thebestsoundperforman e wasobtained usingmetalsponges
withopen- ell stru ture(Fig. 2.26).
The losed- ellmetalli foamsaretoostito onvertsoundenergyintoheatbyvibrating
of their ell walls. However, thesound absorption performan e of losed- ell aluminum
foams an be signi antly improved by optimal opening of losed- ell stru ture using
pro esseslikeholedrilling,rolling and ompression(Fig. 2.27). Inordertomaintainthe
bestabsorptionvaluesofthe hosenmaterials, theair hannelsshouldallbeopento the
surfa e sothat sound waves an propagate into the material. If pores are losed, as in
losed- ell foam, thematerialisgenerally apoor absorber.
Sound absorption materials ontrol airborne noise by redu ing the ree tion of sound
from the surfa e boundaries thus redu ing theoverall noise levels. Sound is attenuated
byvibrationandfri tionlossesbytherepeatedree tionswithinthe ellstru turewhere
full absorption ispossible. Sound absorption properties of foamshowever an bevaried
su h that foams an only be used to eliminate ertain frequen ies. Metalli foams also
an be usedin a wide variety of noise ontrol treatments su h asma hinery en losures,
Figure 2.26: Sound absorption oe ient of various types of ellular Al foams in
om-parison with ber lasssoundabsorbingmaterial [3℄.
Figure2.27: Soundabsorption oe ient asfun tionofopenedsurfa e areaforAlulight
foam (8.9 mm thi k AlSi12 foam, density 580kg/m
3
, holes drilled through the sample,