Effect of Process Parameters on Porosity in Aluminum Lost Foam Process

J. Mater. Sci. Technol, Vol.21 No.5, 2005681Effect of Process Parameters on Porosityin Aluminum Lost Foam ProcessKiyoung KIM2)t and Kyongwhoan LEE2)1) Department of Materials Engineering, Korea University of Technology and Education, 307 GajeonRi, Byungchunmyon,Cheoan, Choongnam, 330-708, Korea2) Advanced Materials Processing Team, Korea Institute of Industrial Technology, 994 Dongchundong, Yunsooku,Incheon, 406-130, Korea[Manuscript received October 13, 2004, in revised form April 22, 2005]Porosity is a main defect in aluminum alloy castings, which is also thought to be severe in aluminum alloy castingsproduced by lost foam process due to the pyrolysis of the polystyrene foam pattern during pouring. Fundamentalexperiments were carried out to evaluate the effect of process parameters such as the melt treatment, the cooling rateand the density of expanded polystyrene (EPS) foam on porosity in A356.2 bar casting. The effect. of melt treatmentincluding degassing and refining was investigated. The effect of cooling rate was also evaluated by changing the moldpacking material such as the silica sand, the zircon sand and the steel shots. Gas entrapment due to the turbulentmetal flow during mold fllig in conventional molding process results in porosity. Mold flling sequence in lost foamprocess is different from that in conventional molding process. The effect of molten metal flow was estimated bycomparing the density of the casting by conventional sodium silicate molding with that by lost foam process. Densitymeasurement was conducted to analyze the extent of porosity in the casting. Source of the porosity in lost foamprocess can be divided into two factors, i.e. turbulence in molten metal flow and entraining residue or gas from thepattern during pouring.KEY WORDS: Lost foam process; Mold packing material; EPS foam; Porosity; Melt treatment1. Introductionpolystyrene foam continues, a bubble of gaseous decom-position product is trapped in the partly frozen casting.Lost foam casting process is still a new technology,Gas entrapment due to the turbulent metal fow dur-and is gaining confidence among manufacturers[1]. It ising mold flling in conventional molding process results inexpected to dramatically improve the competitiveness of porosity. Mold flling sequence in lost foam process is dif-the foundry industry. Many advantages like eliminatingferent from that in conventional molding process. Conse-machining steps, making complex casting without coresquently porosity in lost foam castings is higher than thatand reducing environmental loads can be offered by thisin normal castings because pattern is burned out duringprocess, because molding binder is not added into thepouring of the molten metal.mold of the lost foam process.Turbulence in molten metal flow and entrainingAluminum alloy castings are widely used in the auto-residue or gas from the pattern due to the pyrolysis ofmobile and aerospace industries, and are replacing heav-the polystyrene foam pattern are thought to be sourcesier forged steel or cast iron for the lighter more fuel-of the pore during pouring in lost foam casting. It isefficient automobiles. Producing defect free aluminumdesirable to understand the source of the pore quantita-castings becomes more important.tively. The effects of several related process parametersThe most common defect is the porosity in aluminumsuch as the melt treatment, the mold packing materialcasting. Porosity has an adverse effect not only on theand the density of EPS foam on the porosity in the lostmechanical properties, but also on the surface qualityfoam casting of aluminum alloy were described.of a cast product2. The polymer foam undergoes py-2. Experimentalrolysis and is replaced by the liquid metal in the lostfoam processl3]. One of the unique problems in the pro-A schematic of the lost foam casting system is shownduction of aluminum lost foam castings is the formationin Fig.1. Flask size is 250 mm in diameter and 300 mmof casting defects associated with the pyrolysis productsin height, and casting size was 30 mmx60 mmx250 mm.from the decomposing foam pattern. Various forms ofElectric vibrator is attached under the bottom of thethe pyrolysis-related defects occurred, including internalfask.porosity, blisters, folds, and a rough, orange-peel surfaceneans of a hot wire cut-defect4.ter fr中国煤化工ceramic tube of 30 mmInternal pores are caused by the liquid polymerin inYHC N M H Gouring sprue. Patterntrapped within the liquid metal. The metal solidi-assembly was coatea by alpping with a commercial re-fies before the liquid polystyrene foam escapes to thefractory coating for lost foam pattern and then it wasmetal coating interface. As the pyrolysis of the liquiddried at room temperature for over 24 h.5~7 cm ofmold packing material was first placed in the bottom of↑Prof., to whom corspondence should be addressed,the fAask to form a base for the pattern. Pattern was putE-mailzisholhtac.kr.682J. Mater. Sci. Technol, Vol.21 No.5, 2005Experiments were conducted using A356.2 aluminumalloy ingot which is widely used for various application.250A charge of 15 kg was melted in a graphite crucible usingelectric resistance furnace. Three kinds of melt treatment, Foamwere carried out, i.e. not treated, degassed, and refinedin addition to degassed. Commercial degassing tabletsVacuumand grain refiner were used. Degassing was conducted一pumpby plunging a degassing tablet to the bottom of the cru-吕cible. Care was taken to choose a proper temperature toavoid violent melt convection. After degassing the melt-Packing materialwas covered with fux and the temperature was increasedto 883 K. Requisite amount of grain refiner (0.1% of theFlask, Vacuummelt weight) was plunged into the melt for grain refining,1301which was held down until all bubbling ceased, and de-Sprue.gassed. Gas in melt was estimated by reduced pressuretest (pressure 6.7 kPa) in each heat. Reduced pressuresample of about 150 g was poured into preheated stain-less steel cup. Melt was held for a while further degassing员and inclusion removal. When the gas content dropped toa desired level, the melt was poured at 983 K. Many setsof castings were made, and vacuum was not applied tothe fask during pouring. Up to six chromel/ alumel ther-mocouples of 0.2 mm in diameter were used to measuretemperatures of the casting and the mold.Specimens were sectioned for density measurement af-Vibratorter cooling. Density of the sample was determined by theconventional Archimedes principle. Area of the pore wasFig.1 Design of the lost foam casting systemmeasured using the image analyzer.3. Results and Discussion2.70、2.653.1 Variation in density of reduced pressure samplesFigure 2 shows the variation of density and cross sec-豆2.60tion appearance of reduced pressure samples with timefor the degassed melt. Density increases gradually after区2.55degassing, and reaches a rather constant level with time.会2.50It can be seen that small pores are dispersed and theupper surface of the sample which contains a few pores5 2.45is not swollen from the cross section of the sample. A2.40visual observation through the vacuum chamber gives a50 100150200 25measure of the gas content in the aluminum melt to someHolding time 1 minextent. The shape of the upper surface and bubbles com-Fig.2 Variation of density and cross section appearance ofing to the surface of the sample indicate the presence ofreduced pressure samples with time for the degassedgas during solidification of cup sample. A convex surfacemeltis indicative of high gas content, and concave is indica-tive of low gas content. The higher the specific gravityon the base, and mold packing material was flld vi-of the sample, the less porosity is present. Melt with thebrating the flask with a frequency of 60 Hz. Measureddensity of over 2.58x 103 kg/ m3 was poured.flask accelerations were 2.7 G in x-direction, 0.8 G in3.2 Density distributions in the castingy-direction and 2.7 G in z-direction attaching the ac-Figure 3 shows the relation between the measuredceleration sensor at determined locations respectively.Five kinds of EPS foam with density of 12.5, 15.0, 17.5,density and pore area measured by image analyzer for20.0, 25.0 kg/m3 were used to investigate the effect ofvarious samples by lost foam casting and conventionalfoam density. Sodium silicate-carbon dioxide mold (CO2sand casting. Scatter of data can be seen, but it is evi-dentmold) was also prepared for comparison.中国煤化工ing ince the poreTo investigate the infuence of cooling rate, threeareac)orosity decreases. Gaskinds of mold packing material were chosen, silica sandconte:HCNMHGatedbythedensityofwith an AFS number of 66, zircon sand with an AFSthe casu1ng.The density distributions along the height of the barnumber of 111, steel shots with an AFS number of 54.castings are shown in Figs.4 and 5. Figure 4 showsThe apparent densities of the compacted bed with sil-the distribution by variation in EPS foam density fromica sand, zircon sand and steel shots were 1.49x10*,12.5 kg/m3 to 25.0 kg/m3 in the case of the silica sand2.54x 10号秒1数理kg/m3, respectively.J. Mater. Sci. Technol, Vol.21 No.5, 2005683255p.1010 mm05-Fig.7 Indentation at the mid part of the casting025582602.022.6Measured density 1 (x10^ kg/m')packing. The density distribution by the mold packingFig.3 Relation between the measured density and pore areamaterial is shown in Fig.5. The highest density is foundmeasured by image analyzer for various samples byat top of the bar, while the lowest density is found atthe mid part or mid-low parts. It may be related closelylost foam casting and conventional sand castingto the cooling rate. Solidification time of the upper partwhere the fast cooling by the edge effect produced largetemperature gradient is short. Figure 6 shows the dis-同tribution of the measured solidification time along theheight of the casting for three mold packing materials. Itcan be seen that the upper part is the fastest and the midpart is slowest, because the gas evolved by the pyrolysisof the polystyrene foam pattern and fow turbulence dur-ing pouring is pushed from the fast solidification regionto the slow region. Density difference of 1.7% to 5% inthe same casting is present.15202530Porosity manifests itself in various forms, i.e. mas-Height of specimen/ cmsive shrinkage cavities, macroporosity, dispersed pores, orFig.4 Density distributions along the height of the bar cast-microporositylOo. Small dispersed pores were observed atings (Silica sand packing and EPS foam density frtomthe high density region, while concentrated pores includ-12.5 kg/m3 to 25.0 kg/m3)ing shrinkage cavities at the low density region. Brinellhardness of the inner region of the mid part where con-centrated pores including shrinkage exist is almost halfof the neighbor as shown in Fig.7.3.3 Efect of EPS foam densityFigure 8 describes the relation between the density ofEPS foam used as a pattern and the density of the cast-ing in the case of degassed melt and silica packed mold.Density of CO2 mold casting is also marked in the figure--Sitealhotsfor comparison. The formation of the casting in the lost2.4020foam process is essentially governed by pyrolysis degra-dation of the polymer pattern at the metal front. Therate of formation and the rate of elimination of polymerFig.5 Density distribution by variation in mold packing ma-degradation products can have a vital role in determiningterial .the casting quality in the lost foam processl7. A mixtureof air and foam pyrolysis products in the mold cavityshould escape as soon as possible before solidifed shellof the casting is formed., 400The formation rate of polymer degradation productsis thought to be increased as the foam density is in-creased. The elimination rate of polymer degradationproducts is almost the same because the same mold pack-ing material was used. Densitv of the lost foam castingshows中国煤化工y of 17.5 kg/m3 anddecreasYH! is different from thegeneralCNM H G is iverey propor10152025tional to the foam density.Not only the formation rate and the elimination rateFig.6 Distributions of the measured solidification timeof polymer degradation products but also the metal frontalong the height of the casting for three mold packingshape has a strong effect on the removal of liquid pyroly-ma万均数据684J. Mater. Sci. Technol, Vol.21 No.5, 2005It is thought that the cause of the peak at EPS foam.70-density of 17.5 kg/m3 is the infuence of the metal frontshape. Additional experiments are needed to verify thisobservation.65 +Source of the pore can be divided into two factors,turbulence in the molten metal flow and entraining gas2.0from the pattern due to the pyrolysis of the polystyrenefoam pattern during pouring in lost foam casting. It canbe thought that“A”in the figure is the contribution of-- -Arcooindflow turbulence on the porosity and“A+B”is the contri-2.505bution of flow turbulence and gas from the pattern dueDensity of EPS foam/ (kg/m)to the pyrolysis of the polystyrene foam pattern duringpouring. Since EPS foam decomposes into hydrogen andFig.8 Relation between the density of EPS foam and thecarbon during pouring, carbon may exist in the pore ofdensity of the casting when the melt is degassed andthe lost foam casting. Figure 9 shows the evidence for thepacking material is the silica sandexistence of carbon in the pore of the lost foam castingby WDS analysis, while carbon was not detected in thepore of conventional CO2 mold.(a)Densities of an air cooled specimen which was solidi-fied in the cup, an CO2 mold specimen and lost foam cast-ing is 2.66x103, 2.64x103, and 2.61~2.63x103 kg/m3,respectively. So the contribution of fow turbulence onthe casting can be estimated about 10 kg/m3 and thecontribution of flow turbulence and gas from the patterndue to the pyrolysis of the polystyrene foam pattern canbe estimated about 30~50 kg/ m3 in density scale.It can be concluded that the porosity in the lostfoam casting is higher than that in CO2 mold casting(b)by 20~40 kg/m3 in density. It suggests that it is dif-ficult to obtain better mechanical properties without aspecial treatment in the lost foam casting than in theconventional casting, since mechanical properties of thealuminum casting is related closely to the porosity.3.4 Efect of mold packing materialsThe variation of casting density with mold packing150 μmmaterials is shown in Fig.10. The melt was degassed andEPS foam densities were 12.5 and 20.0 kg/m3. DensityFig.9 WDS result for carbon detection in the pore: (a) SEMof the CO2 mold casting is also marked in the figure forcomparison. Density of the steel shots packed casting isimage, (b) carbon distribution (white dots)higher than the other two cases, however lower than thatof CO2 molding. Mold flling time for lost foam casting70厂was about 8 s, while the molten metal was flled into CO2mold within 3 s, because the presence of the polymer dur-2865-ing mold flling acts as a resistance to the flow of liquid:。日metal.260Metal velocity at the ingate was 0.018 m/s in thisexperiment, which has a signifcant effect on the cast-ing quality, primarily because the rate of foam replace-厂。 125kgm^ment affect the type and numbers of defects formed.0.015~0.02 m/s is recommended8l generally that result50LCO2 Silca sand Zrsand Steel shotsin a minimum number of defects such as misrun, surfaceFig.10 Variation of casting density with mold packing matecollapse, porosity, blister, folds and penetration in lostfoam casting.rialsFioure 11 chowe the variatin of cooling capacity withsis products from the kinetic zone and defect formation inmold中国煤化工es of the silica sand werelost foam casting. In the case of a flat or concave metalchosd|H2g capacity of the moldis esN MH G time or dendrite armfront profile, the liquid pyrolysis products tend to ac-cumulate on the metal front and may get trapped in thespacing of the casting or the heat diffusivity of the moldfreezing metal forming liquid residue related defects, suchpacking material9] or square root of specific heat per unitas laps, folds, and blisters. The bullet shaped (convex)volume in the lost foam casting in this experiment. Theis preferable to eliminate liquid residue related defectsl3).formation rate of polymer degradation products is thou-J. Mater. Sci. Technol, Vol.21 No.5, 2005685casting is much lower.The formation of porosity is also controlled by grainrefining in aluminum casting. Among several advantagesof grain refining, the primary one is an improvement inthe amount and distribution of porosity and shrinkage in2talloy that tends to form microporosityl10. Refiner makespores finely dispersed, but does not affect the amountof formation of polymer degradation products. It seems0- - Specifl heatthat refining the melt does not have any effect on the|+ Soldifcation imecasting density.Silica sandZircon sand Steel shots4. ConclusionsFig.11 Variation of cooling capacity with mold packing materialsThe effects of several related process parameters suchas the melt treatment, the mold packing material and.70 rthe density of EPS foam on the porosity in the lost foamcasting of aluminum alloy were investigated. The high-est density was found at top of the bar, while the lowest.656density was found at the mid part or mid-low parts in a2.608bottom-gated casting of the aluminum bar.There was no proportional relation between the cast-ing density and the EPS foam density. Source of the pore255can be divided into two factors: turbulence in the molten喜:司metal flow and entraining gas from the pattern due to the.50 4No Treated DegassedRefinedpyrolysis of the polystyrene foam pattern during pouringin lost foam casting. The contribution of flow turbulenceFig.12 Variation of casting density with melt treatmenton the density of the casting could be estimated about10 kg/m3 and the contribution of fow turbulence and gasght to be almost the same because of the same moldfrom the pattern due to the pyrolysis of the polystyrenepacking material in this experiment. Therefore the elim-foam pattern could be estimated about 30~50 kg/m3 inination rate of polymer degradation products is seemeddensity.to be closely related to the pore formation. Density ofIt seemed that refining the melt has no effect on thethe steel shots packed casting should be lower from thecasting density because the refiner made the pores finelyviewpoint of solidification rate, since solidification is thedispersed, but does not affect the amount of porosity.fastest in the steel shots packed casting.However the escape path of the gas evolved duringpouring seemns to be more important to remove the poros-Acknowledgementsity. It depends on the various factors such as the moldThe authors thank K.W.Lee and Y .S.Kim for their con-packing ratio, particle size and shape of the packing ma-tribution to this paper. This work was supported financiallyterial. The apparent densities of the compacted bed withby Cleaner Production Technology Development Scheme ofsilica sand, zircon sand and steel shots are 1 .49x 103 ,Korean Ministry of Commerce, Industry and Energy.2.54x 103, 4.1x 103 kg/m3, respectively. Converting thesedensities into the mold packing ratio in volume percentREFERENCESby comparing them with their solid density, 57 .0%, 60.1%and 55.6% are obtained respectively. Average diameters1 ] J.Pletka and J.Drelich: Miner. Metall. Proc, 2002, 19,of the silica sand, zircon sand and steel shots are 0.25, 0.125.and 0.3 mm. Shape of the silica sand is almost round and[2 ] X.Chen and J.E.Gruzlesk: Int. J. Cast Metals Res, 1996,9, 17.that of steel shots is round. On the other hand, the zircon[3] T.V.Molibog and HLittleton: Proc. of Lost Foam Cast-sand is composed of many angular particles. It suggestsing, AFS, Dallas, Texas, 2001, 71.that packing density of zircon sand is high and size of[4 ] S.Bennett, T.Moody, A. Vrieze, M.Jackson, D.R.Askelandthe void between zircon sand grains is small. It can beand C.W.Ramsay: AFS Trans., 2000, 108, 795.concluded that the escape path in steel shots packing is[5] K.Y.Kim, K.W.Lee and K.HRim: J. KFS, 2002, 22, 137.wider than that in the other packing materials, and this[6] Q.Chen and C.Ravindran: J. Mater. Eng. Perf, 2000, 9,results in higher density in the steel shots packed casting.386Figure 12 shows the variation of casting density with[7]中国煤化工Sci. Tecbnol, 1997, 13,melt treatment. The melt was degassed and EPS foam841densities was 20.0 kg/ m3. Density of the CO2 mold cast-[8] C.HMYHCN MH G and PStroom: Proc.ing is also marked in the figure for comparison. Densityof Lost Foam Casting, AFS, Dallas, Texas, 2001, 105.of the refined melt casting is almost the same as that [9 ] S.Koroyasu and M.Matsuda: J. JFS, 1996, 68, 949.of degassed melt casting, while that of no treated melt10] A.Samuel and F.Samuel: AFS Trans., 1992, 100, 657.