Вы находитесь на странице: 1из 6

Materials Science and Engineering A 373 (2004) 309314

Homogenization of ductile iron using partial melting aided by modeling


M. Nili-Ahmadabadi , M. Mosallaiee-Pour
Department of Metallurgy and Material Engineering, Faculty of Engineering, University of Tehran, P.O. Box 11365-4563, Tehran, Iran Received 6 August 2003; received in revised form 28 January 2004

Abstract To provide sufcient hardenability, ductile iron is alloyed with some elements like Mn, Cu, Mo or Ni. However, alloying elements segregation is intrinsic to solidication of ductile iron. Segregation of these elements causes a non-uniform structure which reduces the mechanical properties of iron. It was shown by many researches that high temperature solid state homogenization even after a long period does not effectively reduce the segregation of alloying elements. In this study, partial melting homogenization (PMH) of 1% Mn ductile iron was used for studying homogenization. The homogenization was carried out at different temperature for different periods. The results show that heating above eutectic temperature increases liquid fraction as a function of time while, heating between eutectic temperatures of different region causes melting of heavily segregated regions only. It was shown that the liqueed region shrinks during PMH process which eventually disappears. SEMEDX analysis results show heating below eutectic temperature reduces segregation extensively, while above it, segregation intensity increases. Using calculated segregation prole, assuming uniform chemical composition of liquid pool and local equilibrium at solidliquid interface, homogenization time for different homogenization temperatures were calculated. The calculated results show good agreement with experimental results. 2004 Elsevier B.V. All rights reserved.
Keywords: Ductile iron; Partial melting homogenization; Segregation; Modeling

1. Introduction Deviation from average chemical composition in solidied alloy is called segregation [1,2]. In ductile iron, graphite promoting elements segregate around graphite nodules while carbide promoting elements segregates in the intercellular regions [35]. Segregation of alloying elements has adverse effect on the mechanical properties of ductile iron and ADI [6,7]. As a consequence of non-homogenized distribution of alloying elements in ductile iron, different regions of iron has different phase diagrams, therefore, austenitization temperature varies from near nodular graphite to intercellular region [3,8]. Segregation of substitutional elements causes carbon segregation in the intercellular region after austenitization which in turn changes kinetics of austempering process in different regions and re

duces heat treatment processing window [3,7]. It was shown that the conventional solid state homogenization does not effectively reduces alloying elements segregation even after a long high-temperature austenitization [811]. Purdy et al. applied partial melting homogenization (PMH) process successfully for homogenization of NiSnCu alloys [12]. In this study, the process was applied on ductile iron for the rst time to improve segregation problem in ductile iron. Special attention was paid on homogenization of high Mn ductile iron and calculation of homogenization period as a function of temperature and ductile iron chemical composition.

2. Process denition PMH process consist of the following stages: 1. Heating to homogenization temperature. 2. Isothermal solidication of liqueed region. 3. Homogenization.

Corresponding author. E-mail address: nili@ut.ac.ir (M. Nili-Ahmadabadi).

0921-5093/$ see front matter 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.msea.2004.02.004

310

M. Nili-Ahmadabadi, M. Mosallaiee-Pour / Materials Science and Engineering A 373 (2004) 309314

Selection of homogenization temperature is critical for successful PMH process. Homogenization temperature should be selected between eutectic temperature of regions near graphite nodules and the intercellular region. It was shown that eutectic temperature of cast iron is a function of alloying elements concentration as shown in the following equation [13]: TE = 1148 + (4 %Si) + (5 %Cu) + (8 %Al) + (4 %Ni) (2 %Mn) (1)

4. Calculation method In this research the following assumptions were made: 1. Unidirectional alloying elements diffusion. 2. Planar and local equilibrium at solidliquid interface. 3. A homogenized liquid pool because of complete liquid convection. As the rst step, alloying elements prole as functions of intergraphite distance of as-cast specimen was calculated using Scheil equation [2] (Eq. (2)): Cs (t) = Cs (0) K (1 fs (t))(1K) (2)

where TE is the eutectic temperature, and %Si, %Cu, %Al, %Ni and %Mn the weight percent of silicon, copper, aluminum, nickel and manganese alloying elements, respectively. The above equation shows that carbide promoting elements like Mn which segregate in the intercellular region reduce eutectic temperature while graphite promoting elements like Si, Ni, etc. which segregate around graphite increase eutectic temperature. Therefore, during PMH process, liquid pools form in the intercellular region, while regions near graphite nodules remain solid. As a consequence of diffusion ux of elements inward and outward to liquid pool, its eutectic temperature increases and liquidsolid interface moves to liquid, which eventually, the whole liquids isothermally solidify and homogenization achieves.

where Cs (t) is the chemical composition of each element in the solid, Cs (0) the nominal composition of alloying element and K the distribution coefcient that is 1.17 for silicon and 0.7 for manganese [5] (KSi = 1.17 and KMn = 0.7). In this model, it was assumed that solidication starts from graphite and nish in the intercellular region which is coincident with experimental results. By calculating prole of alloying elements in each point of intergraphite region and using Eq. (1), eutectic temperature was calculated as a function of intergraphite distance. Fig. 1 schematically shows the procedure used for calculation of chemical composition of each node during PMH process at time step of t + 1. The nite difference method used in this study according to the following equations: C Ct+1,i Ct,i = t t (3) (4) (5)

3. Experimental procedure Ductile iron of composition given in Table 1 was prepared. Melting was performed in a medium frequency induction furnace. The base iron was treated with a 2.5 wt.% MgFeSi alloy for spheroidization followed by post inoculation with Fe75% Si. The 15 mm 15 mm 15 mm specimens were heat treated in an atmosphere controlled furnace. Homogenization temperatures were selected according to calculated eutectic temperatures such that regions around graphite nodules remain solid and liquid pools form in the intercellular regions [1315]. Optical and SEM microscopies were used to study microstructure of specimens. To measure chemical composition prole of each element as a function of intergraphite distance, the average value of over 20 Spot-EDX measurement of equivalent intergraphite distance was assumed as chemical composition of that region. The volume fraction of liquid pools was measured using image analysis.
Table 1 Chemical composition of casting Fe C Si Mn Mg Remaining 3.4 2.5 1.0 0.035

Ct,i+1 2Ct,i + Ct,i1 2 C = 2 x x2 Ct+1,i = Ct,i + D t x2 (Ct,i+1 2Ct,i + Ct,i1 )

where C is alloying element local composition, t the time step, x the mesh size, Ct+1,i the concentration in node i at t + 1 time step, and D the diffusion coefcient [16]. It was assumed that D is constant in the experimental temperature. According to Fig. 1 and Eq. (5), alloying elements concentration in node i at t + 1 time step is calculated with the concentration of i + 1, i and i 1 nodes at time step t. To converge the calculation results, t and x was selected such that the following equation satises: D t 1 2 x2 (6)

Fig. 1. FDM mesh for numerical solving.

M. Nili-Ahmadabadi, M. Mosallaiee-Pour / Materials Science and Engineering A 373 (2004) 309314

311

Fig. 2. Plane of solid and liquid pool between two nodular graphite during homogenization. Fig. 4. Liquid pool solidication during PMH.

Fig. 2 schematically shows solidliquid interface. As the gure shows during PMH, liquid pool forms in the intercellular region because of higher Mn content, while region near graphite nodule remains solid due to higher Si content. Due to symmetry, calculation had been carried out for half of intergraphite space. According to Fig. 3, the amount of Mn in the liquid pool at time t can be calculated as follows:
1 CL,Mn (t) 2 (WL (t))

as follows: CL,Mn (t) (WL (t)/2) CS,Mn (t + 1, i) [(WL (t) WL (t + 1))/2] DMn (CS,Mn /x) t CL,Mn (t + 1) = WL (t + 1)/2 (9) Mn concentration in the solid adjacent to the liquid pool is calculated by assuming interface local equilibrium: CS,Mn (t + 1, i) = KMn CL,Mn (t + 1) (10)

(7)

where CL,Mn (t) is Mn concentration in liquid pool at time step t and WL (t)/2 is half width of liquid pool. At next time step (t + 1 time step), amount of Mn in the previous liquid pool (at time step t) is:
1 CL,Mn (t + 1) 2 (WL (t + 1)) + CS,Mn (t + 1, i) 1 2 (WL (t) WL (t + 1))

where K is Mn distribution coefcient. In as much as solidliquid interface moving to the liquid pool (liquid part shrinks), the liquid width was calculated in each step and solid region was re-meshed: WL (t + 1) = WL (t) L (11)

(8)

where CS,Mn (t + 1, i) is Mn concentration in i node of solid that solidies between time interval of t and t + 1. The rst part of Eq. (8) is the amount of Mn in the remaining liquid pool at t + 1 and the second part is the amount of Mn in the solid that formed between time interval of t and t + 1. Using Ficks rst law; Eqs. (7) and (8), the Mn concentration in the liquid pool at time interval t + 1 can be written

where L is liquid volume reduction in each time step of homogenization. Liquid pool was calculated by using lever rule at each homogenization temperature and solidication time was assumed constant. Fig. 4 shows that the solidication time is not a function of homogenization temperature and is almost constant. Although liquid pool at higher temperature has larger size, it can be postulated that at higher temperature diffusion rate is higher which consequently different size liquid pool solidify with in the same period. The same result was expressed by Purdy et al. [12]. Mn concentration in each solid mesh was calculated by solving diffusion equation numerically as follows:
Mn Ct+1,j =

DMn t Mn Mn Mn Mn (Ct,j+1 2Ct,j + Ct,j1 ) + Ct,j x2

(12)

Mn where Ct,j is Mn concentration in jth mesh at t time step. The same equations were used for Si by knowing that Si ux is from solid to liquid.

5. Results Fig. 5 shows the calculated result of alloying elements prole of as-cast specimen as a function of intergraphite distance. This gure shows that Si segregates around graphite

Fig. 3. Schematic prole of alloying element variation in the intergraphite space as a function of homogenization period.

312

M. Nili-Ahmadabadi, M. Mosallaiee-Pour / Materials Science and Engineering A 373 (2004) 309314

Fig. 5. Calculated alloying element prole as a function of intergraphite distance.

Fig. 9. Effect of homogenization period on the measured Mn prole as a function intergraphite distance.

Fig. 6. Calculated eutectic temperature as a function of intergraphite distance.

Fig. 7. Microstructure of the as-cast specimen heated above the eutectic temperature (L: ledburite phase).

and Mn concentrated in the intercellular region. The calculated result coincides with experimental results reported by many researchers [9,17,18]. Fig. 6 shows the calculated eutectic temperature as a function of intergraphite distance. The gure shows that near graphite where Si segregates eutectic temperature is higher than the intercellular region where Mn segregates. Therefore, it is expected that heating of specimen to temperature just above eutectic temperature of intercellular region lead to the formation of liquid pool in this region. This concept was conrmed by the microstructure shown in Fig. 7. The gure shows microstructure of specimen heated to a temperature above eutectic temperature of intercellular region. The gure shows that liquid has been formed in the intercellular region which upon cooling to room temperature transformed to ledburite, while the region around graphite nodule has been left solid. The calculated variation of Mn and Si concentration in the liquid pool and solid region is shown in Fig. 8. The gure shows that after 50 min homogenization, Mn concentration in the liquid pool decreases while Si concentration increases. The situation in the solid region is on the contrary. Si concentration reduces while Mn concentration increases. In addition, this gure shows that during homogenization, solidliquid interface moves to the liquid and eventually the liquid pool solidies completely. Variation of measured Mn concentration as a function of homogenization period is shown in Fig. 9. The gure shows

Fig. 8. The calculated variation of Mn and Si concentration in the solid region and liquid pool as a function of homogenization period. The gure show the liquidsolid interface move to the liquid.

M. Nili-Ahmadabadi, M. Mosallaiee-Pour / Materials Science and Engineering A 373 (2004) 309314

313

Fig. 10. Measured concentration prole of Mn and Si in the as-cast and homogenized specimens using PMH and solid state homogenization.

that Mn concentration in the intercellular region reduce to about 1.7 wt.% of after 1 h homogenization, while the as-cast composition in the same region was about 2.2 wt.% (Fig. 5). Increase of homogenization to 5 h reduces Mn concentration of intercellular area to about 1.2 wt.% and increase Mn concentration near graphite nodular to about 1.1 wt.%, which indicates the effectiveness of PMH process. This matter can be well seem by examination of Fig. 10. The gure shows how after 5 h partial melting homogenization process; Mn and Si concentration prole is rather uniform when compared with chemical composition of the as-cast specimens. In the gure, the results of solid state homogenization at 980 C for 10 h [8] is also added for a better comparison. The gure shows that homogenization at solid state does not effectively change alloying elements segregation while PMH process after a shorter period leads to an almost uniform alloying elements concentration. The calculated and experimental results of Mn and Si concentration of homogenized specimen as a function of intergraphite distance are shown in Fig. 11. The results show a good agreement between experimental and calculated values. The minor difference is likely due to the calculated prole of alloying element in as-cast specimens.

6. Conclusions The partial melting homogenization of high Mn ductile iron aided by experimental and calculation leads to the following conclusions: 1. Heating above eutectic temperature of intergraphite regions leads to formation of liquid pool in these regions. Diffusion ux of Mn from and Si to liquid pool increases eutectic temperature of the pool so that it eventually solidies. 2. Experimental results show that PMH process successfully homogenized the ductile iron matrix. 3. The calculated results show good agreements with the experimental values of PMH process.

References
[1] D.A. Porter, K.E. Easterling, Phase Transformations in Metals and Alloys, second ed., Chapman & Hall, London, 1992. [2] M.C. Felemings, Solidication Processing, McGraw-Hill, New York, 1974. [3] B. Kovacs, The Effects of Alloying Elements and Their Segregation in ADI, Processing of the International Conference on Austempered Ductile Iron, vol. 1, New York, USA, 1991, pp. 241252. [4] E. Dorazil, T. Podrabsky, J. Svejcar, Micro-inhomogeneity of low-alloy austempered ductile cast iron matrix, AFS Trans. 13 (1990) 765773. [5] M. Nili-Ahmadabadi, Structural control of 1% Mn ductile cast iron aided by modeling of solidication and austempering processes, Ph.D. thesis, Tohoku University, Japan, 1994. [6] J.R. Davis (Ed.), ASM Specialty Handbook Cast Irons, ASM International Publication, New York, 1996, pp. 192204. [7] B.Y. Lin, E.T. Chen, T.S. Lei, The effect of segregation on the austemper transformation and toughness of ductile irons, J. Mater. Eng. Perform. 7 (3) (1998) 407419. [8] A. Owhadi, J. Hedjazi, P. Davami, M. Fazli, J.M. Shabestari, Microsegregation of manganese and silicon in high manganese ductile iron, Mater. Sci. Technol. 13 (1997) 813817. [9] E. Dorazil, High Strength Austempered Ductile Cast Iran, rst ed., Ellis Horwood, London, 1991. [10] G. Jolley, L.M. Grad, G.N.J. Gilbert, C. Eng, G. Mesh, Segregation in Nodular Iron and its Inuence on Mechanical Properties, The British Foundryman, March 1967, pp. 7297.

Fig. 11. The calculated and experimental prole alloying element as a function of intergraphite region.

314

M. Nili-Ahmadabadi, M. Mosallaiee-Pour / Materials Science and Engineering A 373 (2004) 309314 [15] M.D. Chaudhari, R.W. Heine, C.R. Loper, Potential applications of cooling curve in ductile iron process control, AFS Trans. 74 (1997) 379386. [16] J. Crank, The Mathematics of Diffusion, second ed., Oxford Science Publication, London, 1989. [17] M. Nili-Ahmadabadi, T. Ohide, E. Niyaa, Effect of successive-stage austempering on the structure and impact strength of high Mn ductile iron, Cast Met. 5 (1992) 6272. [18] M. Nili-Ahmadabadi, E. Niyama, J. Echigoya, TEM study of high temperature bainitic transformation in 1% Mn ductile iron, Mater. Sci. Eng. A 194 (1995) 8798.

[11] J. Lacaze, A. Boudot, V. Gerval, D. Oquab, H. Santos, The role of manganese and copper in the eutectoid transformation of spheroidal graphite cast iron, Metall. Mater. Trans. A 28 (1997) 20152025. [12] G.R. Purdy, D.V. Malakhov, A. Guha, Homogenization of multicomponent alloy via partial melting, J. Phase Equilib. 22 (4) (2001) 439450. [13] Metals Handbook, vol. 15, ASM International Publication, 1990, pp. 168185. [14] R.W. Heine, Austenite liquidus, carbide eutectic and under cooling in process control of ductile base iron, AFS Trans. 59 (1995) 199 205.

Вам также может понравиться