You are on page 1of 8

T H E EFFECT OF ALLZ1I\ILM C O U T E h T AND GRAIN REFINEMENT

OU POROSITY FORMATIOh I\ Mg-AI ALLOYS

Paul L. Schaffer, Young C Lee and Arne K . Dahle

C R C for Cast Metals Manufacturing (CAST) Department of hlining, Minerals and Materials Engineering The Lnikersity of Queensland Brisbane Qld 4072, Australia

Abstract Porosity is detrimental to the mechanical properties surface finish and pressure tightness of castings and i t is therefoie iiiiportdnt to understand the niechanisnis that contiol porosity formation 4 significant amount of research has been perfoinied on the relationship betueen composition and porosity formation i n aluniinium alloys, hoMe\er little mark has been performed on magnesium-based alloys Size and morphology of priniar) phase and eutectic, permeability and solidification range are influenced b> alloy composition and grain refinement and their impact on porosity forination has been studied in the present \\ark Castings of \ar)ing aluniinium content from pure magnesium to Mg-330/0ut Al dllo) u e r e produced and the sample density Mas anallsed using Archimedes principle to deteriiiinc the effect of alloying content on porosit? formation 1he porosity location dnd morphology \\as then characterised by optical iiiictoscopy Alloys \sith high porosit) lelels \+ere then grain refined by the carbon inoculation method to inhestigate the etfect of grain s i x on porosity Introduction The use of niagiicsiuni in commercial applications has increased dramatically in recent years, and growth has been projected to be around 12% per annuni for the next decade [I].The increased deiiiand can be attributed to magnesium's excellent strength to \\eight ratio. making i t particularly attractive for automoti1.e ~ipplications. hlagnesium components are currently mainly produced by casting, and high pressure die casting is the dominant casting method Other casting methods such as permanent mould sand and lo\\-pressure die casting are onl? used to a limited extent but are likely to become more important The increased use of magnesium components has brought nith i t the need for higher integrity castings, defect minimisation and microstructure optimisation Porosity is undesirable in magnesium alloy castings as i t leads to reduced tensile strength, elongation and fatigue life [2,3] In order to meet the deniand for high quality castings, a detailed understanding of porosity formation in magnesiuni alloys and the contributing parameters inust be realised in order to de\elop prehentice measures and optimise the performance and properties o f cast niagnesium components Gencrally, t\vo main factors control the tendency of a casting to form porosity upon solidification [4] The first factor I S the amount of dissol\ed gas particularly hydrogen, in the melt \shich is segregated to the liquid and subsequently nucleates gas pores upon solidification A considerable aniount o f Mark has been conducted to clarify the nucleation of pores on solidification o f various aluminiuni alloys I t has been reported that heterogeneous nucleation of the pores on inclusions such as oxide films [ 5 ] as \\ell as Larious particles present in the melt [6] plays a dominant role as a mechanism for porosity formation This mechanism is strongly influenced bq the amount of turbulence during mould filling, Mhich can also lead to entrained air pochets in the casting

Magnesium T e c h n o l o g 2001 Edited b> J Hr!n TMC (The Minerals. Metals Br Materials Societ)). 2001
87

Thc second contributing factor to poi-osity formation in a casting is solidification shrinkage which is relatcd to the volumetric chaiigc on transformation froin liquid to solid. This is important because i t initiates 3 hydrostatic pressure gradicnt within the casting that subsequently drives the fccding of liquid through the mushy zone during solidification. On frewing, most alloys (except for skin freezing alloys such as cutcctic alloqs) undergo a stage where both solid and liquid coexist (mush). A s solidification progresses. the fraction solid reaches a point \\here equiaxed dendrites start to impinge on one another, ie. the dendrite coherency point [ 7 ] .Recent work has sho\\n that the dendrites arc still free to m o \ e relative to each othcr and first interlock at a point later in the solidification pi-occss callcd the niaxiiiium packing fraction \\here a rigid dendritic network is formed [S]. After niaximuni packing, fccding becomes difficult, since interdcndritic feeding bcconies the dominating feeding mechanism and the reduced permeability of the niushy Lone restricts efficient flow. The potential for porosity formation is high during this stage because the iiiterdendritic pernicability decreases with increasing solid fraction as the feeding channels constrict.

in magnesium alloys. Gruzleski et al. [12] reported that a small addition of strontium to A291 alloy significantly reduced the tendency of porosity formation and caused a finer grain size, but a clear explanation of the results was not g i w n . This lack of hnowledge initiated the present study to obtain a better understanding of the effect of aluminium content and grain s i x on the formation of porosity in magnesium alloys.

i Campbell [4] suniniarised the fw characteristic feeding iiicchanisnis that can occur during the solidification of alloys. Figure I . .4tiiOng these mechanisms, liquid and mass feeding are relati\rly unconiplicated because of the lo\v \.iscosity and wide actile feeding path. These are the major feeding nicchanisnis up to the point of maxiniuiii packing, mass feeding increasing rclatiw to liquid feeding with increasing solid fraction. Beyond and tortuosity of the interdendritic channels, packing, the s i ~ c the presence and fraction of eutectic, thc rigidity of thc dendrite net\iorh and its pcriiicability all affect the feeding efficiency and the dri\ing force for porosity formation.
Clelt treatment, particularly grain refincnicnt, is being practiced to i m p r o x niechanical properties, not only through production of a finer p a i n si/e, but also through niininiisation of pore size

Figure 1 : Schematic illustration of feeding niechanisnis and the pressure gradient caused by solidification shrinkage. The pressure drops as the solid fraction increases.

Expmincntal Procedure

Casting and Grain Refinement Each alloy \\as prepaied in a mild steel crucible coated n i t h boron nitride The al1oSs \bere produced i n an elcctiical resistance furnace using commercial punt) ingots ( 9 9 7 \ i t o " ) of magnesium and aluniinium Clclting and casting \\as conducted under a protectiLe atmosphere of 0 5 % SF, in dry air in order to pre\ent sebere oxidation Each alloy Mas melted and heated to a temperature of approximately 720C foi 20 minutes and e a t i n g was conducted at a temperature 70C a b o \ e the equilibrium Iiquidus teniperaturc of the alloys The mclt Mas poured into a steel mould. preheated to 200"C, Nith a cylindrical cabity \\ith diameter of 50 mm, height of 50 mm and a ibdll-thickness of 2 inni The niould was placed on an insulating felt and allo\+cd to cool freely n i t h no thermal insulation of thc top surface Grain refinement \\as conducted for alloys n i t h 9 \ i t n o 1 I \ \ t o o and 13 u t % aluiiiinium using a grain refiner consisting of I!, \\t% paraffin Max, 76 \\t% CaF,, and 5 \\t% carbon Aftci the grain refiner addition (addition of 0 3 \ito/o of the grain refiner), the melt \\as held at 750C and stirred for 70 minutes to ensurc sufficient dissolution of the grain refiner Ransleq moulds n e r c also cast froni all the melts to detetniine the theoretical base-line density of the alloy for both sciics of caitings, ie unrefined and grain refined The grain si/c of the samples \\as measuied by the linear intercept method described in ASTM El 12-88, and more than 50 intercepts Mere counted for each grain w e determination

and alteration of pore distribution [9]. Despite a considerable amount of research to inLestigate the effect of grain refinement on casting dcfccts [9-I I]. its effects on the forniation and distribution of porosity is yet to be clarified. It has becn reported that grain refinement in AI-Si alloys produced castings with a finer pore s i x , \vcll distributed in the casting [9]. It has becn sho\+n that grain refinement shifts both coherency and maxiiiiuni packing to higher solid fractions [7,S]. Considering the cffccti\e feeding ability of liquid and mass fccding, i t is therefore generally expected that grain refinement should incrcase the soundness of the castings by extending the ranpe of liquid and mass feeding. Howc\er, the impact of grain refinement on interdcndritic feeding niay be more critical for the formation ofporosity, nhich niay be the reason ivhy there is still a debate on the rolc of g a i n refinement i n porosity forniation. Although the essential mechanisms of porosity formation are
siiiiilar in magnesium alloys as in other alloys, such as aluminium allo)s, there are also significant differences. The thermal properties of niapncsiuni are significantly different from

aluminium. and magnesium alloqs generally display a niuch e finer grain s i ~ than aluminium for identical casting conditions. Hydrogcn is thc main gas contributing to porosity in both iiictals. but the difference in hqdrogeii solubility between solid and liquid is not as large in magnesium as in aluminium. Feu studies h a x been conducted to characterise porosit) foriiiation

88

D c n s i t v Mcas u re men t s Densitq measurements Mere performed by b c h i m e d e s principle, according to the ASTM standard [13] in order to determine thc lebel of porosity i n each sample The theoretical density \\as obtained using the same technique on the Ransley niould castings The lebel of porosity in the sample LLas then cdiculated using these trio densities

Results and Discussion The porosity of each casting obtained from the dcnsiry measurements i s s h o n n as a function of aluminiutn content in Figure 2 It can be readily observed that porosity content is strongly dependent on the aluminium content h o porosity is found in pure magnesium or in the eutectic alloy (33 \Vt% Al) Porosity increases M ith increasing aluminium content from pure magnesium, and a peak in porosity is found at about 1 1 \\t% aluminium Porosity then decreases gradually until a pore-free casting is produced v+ith the eutectic allo) To Lcrify the trends o b s e n e d in Figure 2, t\\o additional samples Mere cast for 9 and 13 u t % aluminium The results from thesc castings arc albo included in the Figure, confirming the obserbed trend The equilibrium freezing range, as defined bq the temperature difference b e m e e n thc liquidus (T,) dnd the solidus (Ts) i n the equilibrium binary Mg-AI phase diagram, h a l e been plottcd along nith the measured porosity content i n Figurc 2 The relationship b e h e e n the equilibrium frcering range and the porosity content is readily obserbed, and the peaks in porositv and maximum freering range are Lery close An incrcased equilibrium freezing range generally correlates M ith increascd porosity, and Lice Lersa A similar result mas dlso reported bq \\ hittcnbergcr and Rhines [I41 M here porosity increased linearly \\ith aluminium content up to the maximum equilibrium freezing range (ie maximum porosity le\el in magnesium with 13 \\t% aluminium alloy) for then to decrease The maximum in porosity correlates M ith the first occurrence of eutectic in the microstructure according to the equilibriuin phase diagram

The sainplcs \rerc sectioned lcngthn ise through the centre of the cklinder One hdlf \\as polished to a I p m surface finish and etched ~ i t phospho-picral etchant (0 7 ml HIPOI, 4-6 g picric h acid. I00 nil ethanol (95%)) to rebeal the as-cdst niicrostructure for metallogrdphic analysis These samples \rere used to characterise the phases present and the location and morphology of the porosity The rcnidining half of the samples \\ere solution treated at I O U aluminium content castings \\ere then ctchcd using acetic-picro acid (10 ml acetic. 4 2 ml picric, 10 ml H 2 0 and 10nil ethyl alcohol), Mhile citric acid ( 5 g citric acid, 100 ml distilled \\ater) Mas used for the medium and high aluiminiut~icontent castings The etched samples s e r e thcn used for niedsurcnients of grain size
413C for 2 hours The

,.

Unrefined Alloys

A Grain Refined Alloys

Equilibrium Solidification Range

F i p r c 2 Porosit) \crsus aluniinium content for samples poured at J constant superheat of 70C Thc equilibnuni frcwing range 2s d function o f aluminium content is also gi\en Note that the niauimuni porosity i a l u e occurs at 1 1 \ 4 t o i O , Mhich is thc maxinium equilibrium aolidification rdngc

89

The low porosity content obtained for pure magnesium and the eutectic alloy ( 3 3 u t % Al) can be readily understood as these alloys would be expected to freeze near isothermally and therefore be so-called skin-freezing. enabling efficient feeding throughout the solidification process. The result that maximum porosity correlates with the maximurn in freezing range and the appearance of the first eutectic with increased aluminium content is not surprising. However, the direct relationship between equilibrium freezing range and porosity content is unexpected since the alloys are not expected to solidify under equilibrium conditions. In other alloy systems the peak in porosity is normally correlated with the appearance of the first riori-eqiciiibrricr,2 eutectic. This is the composition where the maximum hydrostatic pressure is developed and therefore causing a maximum in porosity. Hot tearing is another casting defect that is related to the maximum freering range and the presence of eutectic liquid. Measurements of hot-tearing susceptibility as a function of aluminium content in Mg-AI alloys sho\v a peak at about 1 u.t% aluminium [ 151. The Schcil equation can be used to predict non-equilibrium solidification; CL = Co(l - Js)k . where CL is the concentration of the solute in the liquid, C0 is the original concentration of the solute elcment and k is the partition coefficient. This equation proposes that eutectic forms at solute levels much lower than those predicted by the equilibrium phase diagram, The non-equilibrium freezing range and the volume fraction of eutectic, calculated by the Scheil equation for the ,Mg-.Al system, are sho\vn in Figure 3.

The maximum non-equilibrium freenng range is obseri ed around 1 u t % Al, which correlates cery \bell \\ith the measurements of hot tearing, but this clearly does not correlate \vith the peak in porosity According to the model for calculating hydrostatic pressure proposed by Campbell [4], the maxinium hydrostatic tension developed by the flow of feed metal through the dendrite network should occur at about I a t % Al, but the porosity results indicate that this is not the case The equilibrium freezing range is a much better correlation to porosity content The maximum porosity occurred at 1 I \\t% Al \\here the freezing range calculated b y the Scheil equation is about 50C less than the m a u m u m at about 1 ut% AI
I t is not entirely clear \\hy a maximum in porosity should correlate \+ith the appearance of the first non-equilibrium eutectic liquid [4] First, the holunie fraction \ \ i l l initially be extreme11 IOR and the eutectic may therefore be located in small, isolated, pockets, x h i c h are not contributing to the operating hydrostatic tension across the mushy zone Open channels where flou is possible only form u h e n a critical Lolume fraction of eutectic is exceeded This means that before the critical composition is reached, the part of the mushy zone where actiLe feeding occurs is closer to the equilibrium solidification range u h e r e most o f the non-equilibrium solidification \\ill also occur Furthermore, according to the Scheil calculation the eutectic colume fraction for a 2 w t 9 o Al alloy, for example is dbout I 2%, u h i c h \\auld not contribute cery much to ocerall lecel of porosity if i t is left unfed I t is therefore not impossible that the \olume fraction of eutectic correlating \\ith a peak in porosity is that \$hich is linked \ + i t h sufficiently large interdendritic channels for f l o ~ be possible to and also of sufficient i o l u m e to

400

0.6
0.5
.-

& 350 I

2 300 m
.$ 250
0

5
0

. s +

0.4

: 200
.L

150
100

5
K 0

D-

50

o
0
5
10

0.0
15 20
Aluminium Content (wt%)

Non-Equilibrium Freezing Range

-Eutectic Volume Fraction

Figure 3 . Yon-equilibrium frecring range and cutcctic \,oluiiie fraction for hlg-Al allo>.scalculated using thc Scheil equation.

90

Figure 4. Micrographs of typical pores observed in the Mg-AI castings. (a) 2 \vtO/o Al, (b) 5 wt% Al, (c) 9 u?% Al, (d) 1 1 wt% Al, ( e ) 13 \vt% 4 , (0 wtO%Al, (9)24 \vW' Al. An increase in volume fraction of eutectic and size of pores with increased aluminium content is 1 17 clearly visible.
91

(a) Mg-2 wt% Al

(b) Mg-9 \Vt% Al

Figurc 5 . Typical porosity distributions o b s e n e d in the castings. The castings with low and very high Al contents had porosity concentrated in a small region close to the top (a); while the castings \vith intermediate aluminium contents (ie. 9 , 1 I, 13 \\to6 A l ) had porosity in a iiiuch larger region (b).

;i110\\~

largc pores to gro\v.. This also foi-ms a link to the foi-iiiation of gas porosity through hydrostatic pressure.

Incrusing the aluminiuni content produced a progressi1.e increase in thc amount of eutectic phase surrounding the pores, as can be observed in Figure 4 a) - g). In the loiv aluminium coiitcnt saniplcs, ie. 2 and 5 \vt% Al. the pores are surrounded cntircly by primary magnesium and no cutcctic can be observed. Thc first cutectic is obscrved in the 9 mt?o Al alloy, and the \.oluiiic fraction of eutcctic increases progressively u.ith increasing aluminium content. As the aluminium contcnt is increascd to 13. 17 and 74 \It% Al, the dendrites arc almost completely surrounded by eutectic and the sire of the pores increases, which is also related to the size of interdendritic feeding channels. The decrease in porosity with increased aluniiniuni contcnt beyond the peak in porosity at 1 I \bt% Al is thcrcfore explained by an increasing volume fraction of isotherinally frcexing cutectic and large interdendritic channel s i x alloning for relatiwly unrcstricted feeding, and also a decreasing f c u i n g range with increascd aluminium content.
I t is worth noting that the cutectic morphology in Mg-Al alloys displays significant changes Mith increascd aluniinium content. S a \ c ct ill. [ 16,171 sho\vcd rccently that a range of different iiiorphologics for-ni in Mg-XI alloys. depending on the alloy contcnt and cooling rate. A fully di\,orccd eutectic is observed at Ion aluminiuni contents and high cooling rates. \\hereas the full) eutcctic composition tends to solidify with a fully laniellar or fibrous sti-ucturc. The micrographs shown in Figure 4 reproduce sonic of thesc obser\.ations, shouing divorced, partially di\,orced and fibrous morphologies, and also some solid-state discontinuous precipitation. These changes in eutectic morphology arc bvorth noting as it is not entirely clear how they may affcct interdendritic feeding. hydrostatic pressure and porosity formation. Shearhouse and Mikucki [IS] have suggcstcd that rejcction of hydrogen froin the eutectic MgiiAI phasc during cutcctic solidification is an essential mechanism for iiiicroporosity formation in alloy AZ9l. The intermetallic has an extrcniely l o w solubility of hydrogen. If this phase does play a critical role in the initiation of porosity in Mg-.A1 alloys, i t is possible that the morphology of the eutectic growth interphase

is also important. The present work does not shed any further light on this hypothesis. 0vrelid et al. [I91 have shomn that the hydrogen solubility of the liquid decreases with increased aluminium content, but magnesium melts are claimed to rarely be saturated with hydrogen and hydrogen content is therefore not expected to affcct the trends observed in the present study.

Grain size is another important parameter that is likely to affcct the feeding efficiency and the level of porosity in castings through its impact on coherency and permeability of the mushy zone. The variation of grain size observed in this study is similar to the results reported by Lee et al. [20]. A s for most solute additions to a pure metal, a dramatic decrease in grain size foIlo\vs the initial addition of aluniiniuni to magnesium The decrcase of grain size continues with increascd aluminium content until a saturation of grain density occurs around about 9 \+t% Al. Further increases in aluminium content rcsults in only marginal reductions of grain size. As can be observed in Table 1, the grain size of the non-grain refined alloys is relati\.cly small, particularly when the aluminium content reaches 9 \vto/b. These variations in grain size with aluminium content in the unrefined alloys are very small compared to those observed in, for example, aluminium alloys. It is not clear whether thcsc differences in grain size are sufficient to cause significant variations of the rheological properties of the mushy zone, ic. dendrite cohcrency and maximum packing point. although some variation is likely.
A study by Easton and StJohn [ I I ] has shown that there is an optimum range of grain size which reduces, or at least balances. the tendency for porosity formation (external shrinkage of a particular casting) and hot tearing in AI-Si alloy castings. This is a very interesting concept and means that thc optimum grain s i x will depend on the casting conditions and geometry as well as alloy composition. Since the initial grain size of the samples used in the present study is relatively small, the niarginal changes in grain size with aluminium additions niay h a w a niininial effect on the porosity compared to other paranictcrs such as \,olume fraction o f eutectic.

92

l'hc effcct o f adding a carbonaceous grain refiner to the alloys containing 9. I 1 and 13 \\to/o \\.as investigated to test the effect of grain refinement on grain size and porosity formation. The measured grain si/es of these alloys are g i w n in Table I , and the porosity contents h a w been included in Figure 3. From the results it can be observed that only the 9 \vt% A l alloy had a significant decrease of p i n s i x (Table I ) , and this decrease in grain s i x \\as accompanied by an increase in porosity (Figure 2 ) . One possible explanation for thc increased porosity level n i t h grain refinement is that the reduction in grain size, until the onset of the maximum porosity level. ad\.erscly affects intei-dcndritic feeding by reducing permeability. Grain refinement of the 1 I and 13 \\t?,6 A1 alloys resulted i n an insigniticant alteration of grain sizc and the change in porosity \\as also insisnificant. The increase in porosity with grain reiinenicnt of the hlg-9 \\to% A1 alloy is probably not sufficient to draw conclusions about the effect of grain refinement although the result does show a similarity to that often found in other nietal alloy systems. Reduction of permeability caused by morc tortuous interdendritie channels and also a reduction in the jl/c and number o f preferential flow paths is one likely c\planation

marginally larger compared to those obserked at lo\\ aluniinitini contents (Figure 4) The pores \\ere, h o x e \ c r . much nioie distributed for the intermediate aluminium contents

Conclusions The porosity characteristics of cdst 'LIg-41 alloys as a function of aluminium content up to the eutectic composition (33 \\too A l ) and grain refinement h a \ e been inbestigated The results shoM that The l a e l of porosity increases \{ith aluniiniuni content up to 1 1 \\to$ \\here a peak in porosity is obscr\cd Porosity decreases \\ith increased aluminium content from 1 1 nt% to the eutectic composition The peak porosity represents the \\orst feeding conditions bq the combined effects froni pdrameters such as a \\ idc niuyh) zone. IOM permeability and a criticdl \olunie fi;iction of unfed eutectic The aluminium content correlating with the t m \ i m u m porosity le\el obserbed in this study docs not correspond directl) \+ith the prediction of the first non-equilibrium eutectic liquid around I \ i t % .4I \ + h e x the maximum hkdrostatic tension should appear Porosit) instead sho\\s quite a good correlation \\ith the cxtcnt of equilibrium freezing range The pore size increases L\ith increasing rlluniiniuni content The porositl distribution changes as the ,Iluiiiinitiiii content increases and two distinctl) different distributions \\ere o b s e n e d .4t both IOU and high aluminium contents the porosity is concentrated in a small region close to the top of the sample, Lihile the porosity is much more distributed at intermediate al u m i n i uni contents Grain refinement of a LIZ-9 \\too 41 a1101 resulted in an increased l e b e l of porosit)

1 .iblc I Gr'iin s i x nicdsurements of unrefined and giain refined ~ l l o s s ith different aluminium contents L\
I

Grain S i x
A 111 n I um Con tent mi
(nt"0)

(Wl)

I
I

h o n grain refined dlloys

Grain refined allo)s

I20

108

82

I7

S9
~

,
24
It
iaii

71

Ackno\\ Icdqments The authors \could like to acknoLtledge financial support from the Cooperati\ e Research Centre for Cast Metals hfanufacturing (CAST) C.4ST \\as established and is funded in part by the Australian GoLernnient's Cooperatibe Research Centres Program

be assumed that the combination of the pxameters outlined a b o l c , freezing range, solume fraction of eutectic, size of interdendritic feeding channels and permeability combine to produce the o b s e n e d peak in porosity at approximately 9 to 1 I u t " o 21 This tendency niay be a function of the cooling rate of the casting. i e the peak in porosity may shift depending on the cooling rate, most likely appearing at loner Al contents as the dcgrec of de\ iation from equilibrium increases
The distribution of porosit) in the casting also Laried sigiificantl)~ isith aluminium content hlost of the porosity \\as concentrated in a region close to the top of the casting for her) IOU and \ e r > high duminiuin contents, Figure 5 a) In alloys M ith intermediate aluminiuni contents. the porosity \\as much more distributed as s h o u n in Figure 5 b) Although the general ticnd is that the pole size in the high Al content allobs w d s l a g e r than for the ION l contents, the pore size for intermediate A aluminium contents, ie 9, I I and 13\\t% Al. L C ~ Sonly

References
1

D hlagers and J Burssels, "Global Outlook on the Lse of Magnesium Diecasting in Automoti\e Application", Magnesium Allo\s and Their Amlications, B L Mordike and K U Kainer e d s , Wolfsburg, Germany,( 1998), 105-1 12

F A Fox, "The Properties of Some MagnesiumAluminium-Zinc Casting Alloys and the Incidence of Microporosity", Journal of the Institute of Cletals, (1945), 415-338

93

3.

I . J. Feinberg and J. D. Grimsly, "Tensile Properties of Microshrinkage Graded AZ-63 Magnesium Alloy", AFS Trans., 66, (1958), 409-414.

18.

J. D . I. Shearouse and B. A. Mikucki, "The Origin of Microporosity in Magnesium Alloy AZ9 I Trans. SAE. 103, (1994), 542-552.
'I:

J.

Campbell: Cnstirips, Oxford, ( 1 99 I ) , 175-240.

Butterworth-Heinmann.

19.

E. Rooy, "Hydrogen: The One -Third Solution", .4FS

Trans., 10 I,( 1 993), 96 1-964.


P. Mohanty, F. Samuel and J. Gruzleski, "Experimental Study on Pore Nucleation by Inclusions in Aluminium Castings", AFS Trans., 103, (1996), 55.5-564. 20.

E. Bvrelid, P . Bakke and T. A . Engh, "Solubility of Hydrogen in Magnesium Alloys", Linht Metals 1997, Proc. Int. Symp. on Light Metals, Canadian Institute of Metals, Ontario, Canada, (I997), 14 I - 154. Y . C. Lee, A. K. Dahle and D . H . StJohn, "The Role of Solute in Grain Refinement of Magnesium", Metallurgical and Materials Transactions A, (2000) in press.

7.

A . K. Dahle and L. Arnberg, "Investigation of the Dendrite Coherency Point in Solidifying AI-Si Foundry Alloys", Proc. 4th International Conference on Aluminuni Allovs, Atlanta, GA, USA. (1994), 9198. D. H. StJohn and A. K. Dahle? "Rheological Behaviour of the Mushy Zone and its Effect on the Formation of Casting Defects During Solidification", Acta Materialia, 47, ( 1 998), 3 1-4 1 K. Tynelius, J . Major and D. Apelian, "A Parametric Study of Microporosity in the A356 Casting Alloy System", AFS Trans.,101, (1993), 401-413. R. Fuoco, E. Correa and M . Bastos, "Effects of Grain Refinement on Feeding Mechanisms in A356 Aluminium Alloy", .4FS Trans., 106, (1998). 401-409.
. I Easton and D . StJohn, "The Effect of Grain I. Refinement on the Formation of Casting Defects in Alloy 356 Castings", International iournal of Cast Metals Research. 12, (2000), 393-408.

8.

9.

10.

II .

12.

J. E. Gruzleski and C. A. .4lirawi, "LON. Porosity, Fine Grain Sized Strontium-Treated Magnesium .4lloy Castings", L'nited States patent N o 5,143,564. US Patent & Trademark Office: McGill University, ( I 992).
ASTM Standard B331-93, "Test Method for Density Determination for Power Metallurgy Material Containing Less than T w o Percent Porosity"; American Standard for Testing and Materials: Phi ladel ph ia, P A , E. J. Whittenberger and F. N . Rhines, "Origin of Porosity in Castings of Magnesium-Aluminum and other .4lloys", Journal of Metals, ( l 9 j 2 ) , 409-420. R . A . Dodd, W. A. Pollard and J. W. Meier, "Hot Tearing of Castings", AFS Trans.,65, (1957), 100-117. M . D. Nave, A. K.Dahle and D. H. StJohn, "Eutectic Growth Morphologies in Magnesium-Aluminium Alloys", Magnesium Technoloev 2000, H. I. Kaplan, J. N. Hryn and B. B. Clow eds., Kashville, Tennessee, USA, (2000), 233-242 M . D. Nave, A . K. Dahle and D. H. StJohn, "Eutectic Growth in Magnesium-Aluminium Alloys". Magnesium Technoloev 2000, H. I. Kaplan, J. N. Hryn and B. B. Clow eds., Nashville, Tennessee, USA, (2000), 243-250
94

13.

14.

15. 16.

17

You might also like