influence of process parameters on …influence of process parameters on macrosegregation of...

99

Int. J. Mech. Eng. & Rob. Res. 2012 K V Sreenivas Rao, 2012

INFLUENCE OF PROCESS PARAMETERS ONMACROSEGREGATION OF DIRECTIONALLY

SOLIDIFIED LEAD - TIN ALLOYS

K V Sreenivas Rao1

*Corresponding Author: K V Sreenivas Rao, [email protected]

Directional solidification experiments were carried out on hypoeutectic and hypereutecticPb-Sn alloys. Different combinations of growth rate V and composition Co. were used to investigatetheir effect on longitudinal macro segregation. Macro segregation along the length of the sampleswas observed in hypoeutectic Pb-Sn alloys whereas no such macro segregation was observedin hypereutectic alloys. The intensity of longitudinal macro segregation was found to increasewith the increase in initial tin content of the alloy, increase in distance from the chill end anddecrease in the solidification rate.

Keywords: Directional solidification, Macro segregation, Growth rate

INTRODUCTIONMacro segregation, the non-uniformmacroscopic distribution of the componentsof an alloy during solidification, is a defect,which can occur in real metal processingsystems. These non uniformities, especiallywhen they form high compositional gradients,can be areas of high stress concentration,which cause cracking when an aluminum orsteel billet is extruded or forged or when anas-cast piece, such as turbine blade, isseverely loaded in service (Krane andIncropera, 1997).

ISSN 2278 – 0149 www.ijmerr.comVol. 1, No. 1, April 2012

© 2012 IJMERR. All Rights Reserved

Int. J. Mech. Eng. & Rob. Res. 2012

1 Professor, Department of Mechanical Engineering, Siddaganga Institute of Technology, Tumkur, Karnataka, India.

The physics of macro segregation formationcan be summarized as follows. Segregationstarts at the microscopic level as solidificationproceeds. During solidification, solute isrejected or depleted continuously from theprecipitated solid and the composition of thesurrounding liquid is consequently affected. Ifsignificant concentration gradients aredeveloped at the interface, the interdendriticliquid can be driven simultaneously by thermaland solutal buoyancy, as well as solidificationcontraction. The induced flow will wash awaythe liquid next to the interface, resulting in

Research Paper

100


segregation at the macroscopic level calledmacro segregation (Stefanescu, 2002).

When a homogeneous, stagnant fluid issubjected to a vertical temperature gradient,the resulting instabilities can give rise toconvective flow patterns. However, if theresulting temperature at a particular locationfalls below the local liquidus temperature it islikely to promote solidification. In case of amulticomponent system a mushy layer ofdendritic crystals may form duringsolidification, the interstices of whichaccommodate the residual solute that isrejected during the process. Thus in additionto the temperature gradient created byexternally imposed boundary conditions as wellas due to the release of the latent heat,concentration gradients are also created bythe rejection of the solute across the solid/liquidinterface. The rejected solute may betransported by diffusion on a local scaleleading to micro segregation and byconvective flow on a larger scale leading tomacro segregation. From the practicalviewpoint, the nature and extent of the macrosegregation will determine the quality of thefinal product and hence forms an importantresearch area.

If the melt is cooled from below, the fluidlayer is always thermally stable, and hence,thermal buoyancy driven flow does not occur.Additionally if the system rejects a denserresidual on solidification, both thermal andcompositional fields are gravitationally stable,and convective motions do not occur(Huppert, 1990). Numerous theoretical,experimental and numerical studies havebeen devoted to the fluid mechanics aspectsof solidification, which is presented as a

comprehensive review by Huppert (Kumaret al., 2003).

In a vertically upward growth, the fluid flowwill be controlled by the density gradient in themelt at the interface. If this density gradient ispositive, i.e., the rejected solute is lighter; thefluid flow effect can be significant. In contrast,when the axial density gradient is negative, i.e.,heavier solute is rejected; fluid flow can stilloccur if any horizontal thermal gradient ispresent. Trivedi et al. (2001) have carried outa detailed numerical analysis of the effect ofradial temperature gradient on the intensity ofconvection. In this research an attempt is madeto examine the longitudinal macro segregationduring upward directional solidification undercontrolled experimental conditions. The effectof solidif ication parameters on macrosegregation is investigated.

EXPERIMENTAL PROCEDURE

Alloy Preparation

Six different alloy compositions of Pb-Sn,three each in hypoeutectic (Pb-30Sn, Pb-40Sn, Pb-50Sn) and hyper eutectic (Pb-70Sn,Pb-80Sn, Pb-90Sn) ranges were preparedby remelting 99.9% Pb and 99.9% Sn usinga high frequency induction furnace. Fullystirred melt was rapidly poured in to the castiron mold and the alloy samples wereobtained from this ingot.

Directional Solidification

The apparatus used for upward directionalsolidification is schematically illustrated inFigure 1a. Figure 1b shows the experimentalsetup. In an ideal directional solidificationsystem, the imposed temperature gradient isperfectly perpendicular to the plane of

101


solidification, which is the melt/crystal interface.Lateral temperature gradients are notdesirable because they cause thermalstresses, which lead to the formation ofvacancy, and dislocation defects in the crystals(Reza, 1994). The furnace temperature isdesigned to increase in the direction oppositeto gravity in order to suppress naturalconvection. The rise in temperature oppositeto gravity causes the density to decrease withincreasing height.

The prepared alloy samples of 10 mmdiameter and 50 mm long were placed in SiO

2

crucibles and melted. The molten sampleswere solidified directionally upwards in theexperimental setup consisting of a heating unit,a temperature control system, a samplemoving system and a heat extraction system.Initially the melt is allowed to reach the selectedtemperature and allowed to solidify from thebottom. The water-cooled copper chill insertedat the bottom of the SiO

2 crucible produced

temperature gradient in the melt. The furnacecan be moved upwards or downwards at aspeed of 4-100 ìm/s by means of a translationdevice consisting of a set of gears and a

Figure 1: Directional Solidification Facility

(a) Line Diagram of the Apparatus (b) Photograph of Experimental Setup

Argon Gas

Guide Rod

Furnace

Crucible

CopperChill

Cooling Water

FurnaceStand

Screw Rod

SpeedChangeWheel

Gear Box Housing

102


motor. The directional solidification wascarried out by raising the furnace assembly atvarious translation speeds with respect to thestationary sample. This avoids convection dueto crucible motion. The furnace and the cruciblearrangements were such that the furnacetranslation speed was equal to the directionalsolidification rate.

RESULTS AND DISCUSSIONThe directionally solidified samples werelongitudinally sectioned and polished for

microscopic observation and examined usingScanning Electron Microscope. The SEMEDX spectrum showed the presence of macrosegregation in hypoeutectic Pb-Sn alloys andthe absence of any such segregation in hypereutectic Pb-Sn alloys. Segregation results forPb-Sn binary alloys are shown in Figures 2, 3and 4 in which Cs/Co is plotted againstdistance from the chill end. Cs is thecomposition at the measured location of thesample and Co is the initial (average)composition of the sample.

Figure 3: Macro Segregation of Hypereutectic Pb-Sn Alloy Directionally Solidifiedat 100 m/s

Distance from Chill End (cm)

0 1 2 3 4 5 6

1.3

1.2

1.1

1.0

0.9

0.8

0.7

Cs/

Co

Pb-70Sn

Pb-80Sn

Pb-90Sn

Figure 2: Macro Segregation of Hypoeutectic Pb-Sn alloy Directionally Solidifiedat 100 m/s


0 1 2 3 4 5 6

1.3

1.2

1.1

1.0

0.9

0.8

0.7

Cs/

Co

Pb-30Sn

Pb-40Sn

Pb-50Sn

103


Figures 2 shows the macro segregationresults of hypoeutectic Pb-Sn alloys. It showsthat Cs/Co value increases with the increasein distance from the chill end, indicating thepresence of macro segregation. Thesolidification begins at the bottom andprogresses upwards. The rejected solute (Sn)at the solidification interface near the bottomgets redistributed in the remaining meltconsequently increasing the tin content in themelt. Therefore the alloy that solidifies at thelater stages will be of higher composition.Thus, as the solidification progresses upwardsthe solute content in the melt progressivelyincreases giving rise to macro segregation inthe longitudinal direction. It is also observedthat the variation in the slopes of the threecurves indicates that the macro segregationincreases as the initial tin content in the alloyincreases.

Figures 3 shows the macro segregationresults of hypereutectic Pb-Sn alloys. It showsthat the Cs/Co value is nearly the same alongthe entire length of the sample. This indicatesthe absence of any significant longitudinal

macro segregation in case of hypereutecticPb-Sn alloys for all the three values of Coconsidered. In the case of hypereutectic Pb-Sn alloys the primary dendrites are of tin andthe rejected solute is lead. The density of leadis greater than that of tin, so the soluteconvection effects are minimized. Thereforethe rejected solute atoms will not be carriedaway from the solid liquid interface, whichminimizes the macro segregation.

Figure 4 shows the dependence of macrosegregation on solidification rate. The intensityof macro segregation increase as thesolidification rate decreases. The longitudinalsegregation was greatest in the slowestvelocity samples. At the slowest velocity of4 m/s, there is sufficient time for transport ofspecies by both diffusion and advection in theliquid.

Macro segregation is associated with theeffect of gravity. The convection driven byvertical solutal gradient leads to macrosegregation. In the case of hypoeutectic Pb-Sn alloys the solute (Sn) is lighter than thesolvent (Pb). As the crystals of the primary

Figure 4: Effect of Growth Rate on Macro Segregation of Directionally SolidifiedPb-40Sn Alloy


0 1 2 3 4 5 6

1.4

1.3

1.2

1.0

0.9

0.8

0.6

Cs/

Co

100 m/s

20 m/s

4 m/s

1.1

0.7

104


dendrite (Pb) form, the solute (Sn) in the meltsurrounding the primary dendrite increasesand the density profile in the interdendriticmelt and in the melt immediately ahead of themushy zone promotes natural convection. Inthe case of hypereutectic Pb-Sn alloys thesolute (Pb) is heavier than the solvent (Sn).As the crystals of primary dendrite (Sn) form,heavier solute (Pb) in the melt surroundingthe primary dendrite and in the meltimmediately ahead of the mushy zoneprevents natural convection. In this case boththe solutal gradient and the thermal gradientstabilize the liquid system. In upwarddirectional solidification thermal gradientopposes the solutal gradient. The thermalRayleigh number (RT) and solutal Rayleighnumber (RC) have been defined in theliterature to describe the onset of thethermosolutal convection in the melt (Heinrichet al., 1989). The two numbers are given as:

RT = gG

TH4/K

T

RC = gG

CH4/D

C

Where is the thermal volume expansioncoefficient, g the gravitational accelerationconstant, G

T the thermal gradient, K

T the

thermal diffusivity, the kinematic viscosity, the solutal volume expansion coefficient, G

C

the solutal gradient, DC the solutal diffusivity

and H a characteristic length (for the dendriticarrays, the primary arm spacing has been usedas the characteristic length). Data listed inTable 1 (Sazarin and Hallawell, 1988) wereused to calculate the R

T and R

C under the

conditions of this research. RT = 4.2 108 (G

T)

H4, RC = 6.8 1013 (G

C) H4, R

C/R

T = 1.61 105

(GC/G

T). The solutal Rayleigh number

determines the intensity of macro segregationwhen the thermal gradient (G

T) is not

excessively large, because the solutal Rayleighnumber is much larger than the thermalRayleigh number.

Figures 5a and 5b show the opticalmicrographs of the hypoeutectic and

Properties Value

Kinematic viscosity of melt () 2.47 10–7 (m2s–1)

Thermal coefficient of volumetricexpansion of melt () 1.15 10–4 (K–1)

Solutal coefficient of volumetricexpansion of melt () 5.20 10–3 (wt%–1)

Thermal diffusivity of melt (KT) 1.08 10–5 (m2s–1)

Solutal diffusivity of melt (GC) 3.00 10–9 (m2s–1)

Acceleration due to gravity (g) 9.80 10 (ms–2)

Table 1: Data Used in the Calculationof Rayleigh Number for Pb-Sn Alloy

Figure 5: Microstructures Alloy SamplesDirectionally Solidified at 100 m/s

(a) Hypoeutectic (Pb-50Sn)

105


Figure 5 (Cont.)

(b) Hypereutectic (Sn-10Pb)

hypereutectic Pb-Sn alloys respectively. Thefigures clearly show that, in case ofhypoeutectic Pb-Sn alloys the primary dendriteis of lead whereas in case of hypereutectic Pb-Sn alloys the primary dendrite is of tin, whichis confirmed by SEM EDX.

CONCLUSIONUpward directional solidification experimentswere carried out using hypoeutectic andhypereutectic Pb-Sn binary alloys with differentcombinations of growth rate and composition.The effect of solute convection on the macrosegregation was investigated. The followingconclusions were drawn from the study

• Convection during upward directionalsolidif ication produces extensivelongitudinal macro segregation.

• When the solute is lighter than the solvent,macro segregation results from theconvection driven by the vertical densitygradient.

• Macro segregation increases as the solutecontent in the alloy increases.

• The intensity of macro segregationincreases as the solidif ication ratedecreases.

• When the solute is heavier than the solvent,macro segregation is not observed due tothe absence of convection.

Acknowledgment: The author acknowledges

with gratitude the kind help of Prof. T S Prasanna

Kumar, Department of Metallurgical and Materials

Engineering, IIT Madras, for extending the

experimental and testing facilities and for useful

technical interaction.

REFERENCES1. Heinrich J C, Felicelli S, Nandapurkar P

and Poirier D R (1989), “ThermosolutalConvection During Dendritic Solidificationof Alloys: Part II. Nonlinear Convection”,Metallurgical and Materials TransactionsB, Vol. 20B, pp. 883-891.

2. Huppert H E (1990), “The Fluid Mechanicsof Solidif ication”, Journal of FluidMechanics, Vol. 212, pp. 209-240

3. Krane M J M and Incropera F P (1997),“Experimental Validation of ContinuumMixture Model for Binary AlloySolidification”, Journal of Heat Transfer,Vol. 119, pp. 783-791.

4. Kumar P, Chakraborty S, Srinivasan K andDutta P (2003), “Studies on TransportPhenomena During Directional

106


Solidification of a Non-Eutectic BinarySolution Cooled from the Top”,Metallurgical and Materials Transactions,34B, Vol. 6, pp. 899-909.

5. Reza Mehrabi M (1994), “ModelingTransport Processes in DirectionalSolidif ication: I. Insatabil it ies andNonlinrar Evolution in Binary Alloys: II.Direct Parallel Solution of PDE’s”, Ph.D.Theses, MIT, US.

6. Sazarin J R and Hallawell A (1988),“Channel Formation in Pb-Sn, Sn-Sb and

Pb-Sn-Sb Alloy Ingots and Comparisonwith the System NH

4Cl-H

2O”, Metallurgical

and Materials Transactions A, Vol. 19A,pp. 1861-1872.

7. Stefanescu D M (2002), “Science andEngineering of Casting Solidification”,Kluwer Publishers.

8. Trivedi R, Mazumder P and Tiwari S N(2002), “The Effect of Convection onDisorder in Primary Cellular DendriticArrays”, Metallurgical and MaterialsTransactions A, Vol. 33A, pp. 3763-3775.

influence of process parameters on …influence of process parameters on macrosegregation of...

Documents