Production of rheocast slurries by partial melting through alternative thermomechanical treatments

Production of rheocast slurries by partial melting through alternative thermomechanical treatments

M. Margarido; M. H. Robert

Mechanical Engineering Faculty, State University of Campinas, Cidade Universitária, 13083-970 Campinas, SP. [e-mail]

ABSTRACT

The production of alloys with rheocast structure, where the primary phase consists of globularized particles, by partial melting, involves grain recrystallization and secondary phases melting. In the process known as SIMA (Strain Induced Melt Activation), the raw material is cold deformed by a two-step rolling, making it little efficient. This work investigates the production of thixotropic material, by alternative thermomechanical treatments, of an Al-3.35wt%Cu alloy. Two treatments, which avoid the hot deformation step of SIMA, are performed: Recrystallization and Partial Melting (RAP) and Overaging (OAT) processes. In the first, the as-cast, dendritic, alloy is cold deformed at ambient temperature and then heated to a constant temperature in the mushy zone, to obtain the rheocast structure. In the second, the alloy undergoes a solution and precipitation heat treatment, before cold-deformation, with the aim of obtaining second phase precipitated particles with size and interparticle spacing favourable to the increasing of the nucleation rate of recrystallized grains, during heating of the alloy. It was found that the two routes produce globularized structures, with grain reduction of 85 % and 90 % for RAP and OAT, respectively, with respect to initial grain sizes. In general, OAT process resulted in more rounded and smaller globules than RAP.

Keywords: Semi-solid alloys, rheocasting, thermomechanical treatments, aluminium alloys

Introduction

The increasing tendency for near-net-shape forming of industrial parts, taking advantage of costs reduction and quality improvement, is turning thixoforming and thixocasting into a viable technology - which includes forging, injection and extrusion moulding (Kenney et al, 1988), and with possibilities in deep drawing (Adamiak & Robert, 1998) - for light alloys, steels and high speed steels.

On the other hand, industrial production of thixotropic materials is limited to the eletromagnetic stirring process, in which the molten alloy undergoes high shear stresses during solidification, resulting in a microstructure of fragmented dendrites and rosette shaped particles, which have to undergo a posterior spheroidization heat treatment. Others proposed processes like mechanical stirring, SCR process and passive stirring were not competitive because of lower productivity or difficulty in process control.

Another approach for the production of rheocast alloys is the solid state route, by partial melting of deformed dendritic solid materials (Robert & Kirkwood, 1988). In the solid state process, the deformed material is heated up so that recrystallization takes place before it reaches the semi-solid temperature range. In this range, melting of secondary phases occurs, which surround the recrystallized grains previously formed, resulting in the expected spheroidized microstructure. However, when heating in excess, coarsening phenomena gives rise to undesirable grain growth. The major advantage of the solid state process is that it doesn’t need expensive equipment to obtain the non-dendritic structure nor posterior globularization heat treatment. In the method patented as SIMA (Young et al, 1983) the cast material is extruded at high temperature to fragment the dendritic structure and then cold deformed above a critical strain, before the heating step.

The purpose of this work is to study the behaviour of an Al-Cu alloy after being submitted to simpler thermomechanical treatments, which involve only one deformation step, and to analyze the influence of initial grain size, ageing conditions and cold work level on the final characteristics of the resultant rheocast. According to Loué &Suéry (1995), the initial grain size shouldn’t affect the morphology and final grain size of rheocast slurries, since enough strain, 25 % true strain, has been supplied. On the other hand, working with Al-4.5wt%Cu alloy, Zoqui & Robert (1998) obtained results that show a dependence between initial and final grain sizes, after 40 % of true strain.

The main control parameter that influences the final microstructure of a rheocast material is thought to be the degree of cold deformation. If a critical value is surpassed, over which there is dislocation density saturation, the morphology and final grain size are independent from strain level (Robert & Kirkwood, 1988).

Experimental Procedures

An aluminium-copper alloy, whose chemical composition is shown in Table 1, with Solidus and Liquidus temperatures of 906 K and 915 K, respectively, determined by DTA, was chilled and sand cast to produce samples with two different initial grain sizes.

Part of them has undergone a solution, at 820 K for 2 h, and precipitation, at 653 K, heat treatment. The coarse grained alloy was precipitation heat treated for 6 h and the fine grained alloy for 25 h. Such parameters were determined according to results obtained in preliminary experiments of recrystallization of overaged samples, to search for the times that provided smaller grain sizes. Figure 1 shows recrystallized grain sizes of coarse and fine grains samples with respect to precipitation time, between 2 and 25 hours of heat treatment.

Samples with dendritic and overaged microstructures were submitted, then, to 45 % and 80 % of true strain by compression and rheocasted by partial melting at the constant temperature of 908 K. There was no holding time at this temperature, so that samples were cooled as soon as they reached it. Average heating rate was .92 K/s. In another series, samples submitted to the same casting, heat-treating and strain conditions where allowed to rheocast during 300 s before cooling. Schemes of the thermomechanical treatments employed are showed in Fig. 2.

Optical and scanning microscopy were used to evaluate results. Micro probe analysis was done in order to obtain the chemical composition of localised regions in as-cast and rheocast structures. Quantitative metallography was applied to measure grain sizes, while shape factor, defined as the factor between the larger and the least dimensions of each particle, and solid fraction were quantified by image analysis.

Results and Discussion

In Fig. 3 the macro and microstructures of sand and chilled cast samples can be seen. A dendritic structure of equiaxed grains is observed, with different grain sizes and secondary spacing, due to the different cooling conditions employed. Their values can be seen in Table 2.

Microssegregation of Cu due to solute rejection during solidification can be observed through microprobe analysis across dendritic arms of the as-cast alloy. Figure 4 shows a SEM micrography of the coarse alloy, showing the points of microanalysis, and solute distribution curves for Cu.

One of the microstructural features of rheocast alloys is the absence of microssegregation, due to higher diffusion rates in the mushy zone. Figure 5 shows Cu content in rheocast structures obtained by RAP and OAT with respect to the distance from grain boundaries (distance = 0). It can be seen the flat profile of the curves inside the grains and a peak at the boundary. The Cu content in solid solution is lower than in the as-cast alloy because of the presence of approximately 25 % of liquid, where the solute solubility is higher. The Cu content in the primary phase of the OAT rheocast is higher than the RAP one (1.9 % against 1.5 % averaged); because of the presence of Cu rich precipitated particles in the Al-a matrix.

RAP Thermomechanical Treatment

Typical microstructures of Al-3.35wt%Cu alloy rheocast samples, obtained by RAP process, are presented in Fig. 6. The upper figures show microstructures from initially refined alloy, and at the left it can be seen the less deformed structure. The primary Al-a phase is no dendritic, with 1,5 %wt Cu in solution and entrapped pools of CuAl₂. Fe and Si, the main impurities, were found to be concentrated in pools and in the second phase with eutectic composition in the interglobular regions.

For the refined alloy, primary phase in the rheocast material is globularized, even for less deformed material, while for coarse grained alloy the primary phase of semi-solid material is not completely rounded, due to longer diffusion distances involved and the short time supplied (no holding time at 908 K was allowed) for coarsening phenomena to be effective. The samples submitted to lower deformation present fragmented dendritic arms, in an initial globularization process, where thicker second phase surrounding the solid particles reveals original grain boundaries. The CuAl₂ pools are mainly present inside the primary phase of the initially refined alloy, since the second phase was more homogeneously distributed in the original dendritic structure. While there is no qualitative evidence of the influence of deformation level on the microstructure, the role of initial grain size is predominant in the globularization process.

The average globule sizes, shape factors and solid fractions of rheocast samples obtained by RAP process, with no holding time at the rheocasting temperature, in all tested conditions, are presented in Table 3. Results show a great reduction in grain/globule size in both thermomechanical treatments, indicating efficient recrystallization. It is observed that grain size of rheocasts obtained from RAP is statistically independent on degree of deformation, but globule size of rheocasts from refined alloy are 39 % smaller than from coarse alloy, showing a big influence of initial grain size on the process.

Results of shape factor show that are limits to use RAP process to obtain rheocast materials with well rounded primary phase. Only the refined alloy was successful with this respect, at any degree of deformation, with a average shape factor of 1.65. In the case of coarse alloy, there wasn’t complete globularization, resulting in bigger shape factors.

As the rheocasting temperature is always the same, the solid fraction is kept constant and in reasonable agreement with calculated value by Scheil equation (70 %). Solid fraction after 300 s of rheocasting treatment is of the order of 72 %, nearer from equilibrium condition given by Scheil equation value. The reduction in solid fraction is due to melting of solute rich regions.

OAT Thermomechanical Treatment

When using overaged starting material, rheocasts with globular primary phase surrounded by liquid of eutectic composition are obtained in all tested conditions, but with more heterogeneity for the coarse alloy, showed in Fig. 7. Again, the upper figures show microstructures of initially refined alloy, and the left for less deformed structures. There is a great amount of CuAl₂ particles inside the globules, which may be attributed to precipitation during cooling after rheocasting treatment, or to precipitate particles from the overaged alloy that were not dissolved during heating. There are some bigger pools, too, mainly in the 45 % deformed and coarse grained alloy, resultant from coalescence phenomena, that prevailed to some extent.

The secondary phase is located either at grain boundaries with eutectic composition and as entrapped liquid, for dendritic alloy, or at grain boundaries and as precipitated particles in the primary phase, in the case of overaged alloy. These CuAl₂ precipitates, confirmed by SEM analysis, are found inside the primary phase except for an annular region surrounding the grain boundaries. It can be observed too that these areas are larger where the boundaries are thicker.

Figure 8 shows a SEM micrograph of a fine grained alloy deformed 80 % where this precipitates depleted area can be seen, which is the result of copper diffusion to grain boundaries. Very similar structure was shown in recent work (Braccini et al, 1998), after holding samples of spray deposited Al-10%Cu in the mushy zone.

Rheocasts produced by OAT process presented globular structures in all cases, with smaller grain sizes when compared to that obtained by RAP in the same conditions. The reductions, if compared to initial grain sizes, were 476 % for refined alloy and 1295 % for coarse alloy. In this process, globule size is dependent either on initial grain size than on degree of deformation. The average globule size undergoes a reduction of 24 % when grain size is decreased from 1290mm to 403mm, and of 14 % when deformation is increased from 45 % to 80 %, with better performance for initially refined alloy (17 %), against 11 % for coarse grained alloy; showing that recrystallization of overaged alloys is more effective with smaller grains. Rrecrystallization stimulated by precipitated particles plays an important role in OAT process at any condition. The best rheocast quality (smaller and more rounded primary phase) was obtained in the most favourable initial conditions: higher strain level, smaller initial grain size and bigger size of CuAl₂ precipitates produced after 25 h of precipitation treatment.

The shape factors of rheocasts by OAT show good globularization for all tested conditions and seem not to depend on degree of deformation, for a given grain size. Figure 8(a) shows the influence of the thermomechanical treatments on average globule diameters of rheocasts. Samples previously submitted to 45 and 80 % of cold-work presented grain size reductions of 13 % and 21 %, respectively, when obtained by OAT with respect to RAP process.

The effect of thermomechanical treatments, taking into account initial grain size variation, can be observed in Fig. 9(b). Only the coarse grained alloy presents significant globule diameter reduction of rheocasts produced by OAT with respect to RAP. This is due to the behaviour of refined alloy submitted to 45 % of cold work, which showed no important changes between the different treatments, probably because recrystallization of overaged alloys is less sensitive to lower degrees of deformation (Humphreys, 1977).

Figure 10 shows the influence of thermomechanical treatments on globule diameter for all analyzed parameters. The best absolute result was obtained with OAT process applied to refined alloy submitted to 80 % of deformation, while the best performance was obtained with coarse grained alloy.

Table 5 shows the average globule size variation of rheocast structures obtained by OAT with respect to RAP rheocast structures, for each experimental condition. It can be seen that the size reduction factors of initially coarse alloy are the same for 45 % and 80 % deformed samples (22% and 26 %), which can be attributed only to the change of thermomechanical treatment.

Conclusions

Both thermomechanical treatments processes yield rheocast structures production by partial melting of an Al-3.35wt%Cu. Reductions in globule size with respect to initial grain size are 700 % and 870 %, averaged, for RAP and OAT respectively. Initial grain size influences the morphology and globule dimensions of the primary phase of rheocasts obtained by both processes. The globule size of rheocasts is independent on degree of deformation in RAP process, but dependent on it in OAT. For this last process, an increase in degree of deformation results in a decrease in globule size. In general, overaged Al-Cu alloys can produce rheocast material (OAT process) with more homogeneous structure and smaller globule size, when comparing with traditional recrystallization route (RAP process). By using OAT, reductions in globules dimensions in the order of 26 % can be achieved, when comparing to structures obtained by RAP. Therefore, previous overaging treatments can, at some extent, avoid the growth of globules during heating.

Acknowledgements

The authors are grateful for financial support from CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico).

Paper accepted: July, 2002

Technical Editor: José Roberto de França Arruda

Presented at COBEM 99 – 15th Brazilian Congress of Mechanical Engineering. 22-26 November 1999, São Paulo. SP. Brazil.

• Braccini, M., Salvo, L., Suéry, M. and Bréchet, Y., 1998, Proceedings of the 5^th International Conference on Semi-solid Processing of Alloys and Composites, June 23-25, Colorado, pp. 371-378.
• Adamiak, M. & Robert, M. H., 1998, Preliminary studies on the suitability of rheocast Al alloys for deep drawing, Proceedings of the 7^th International Scientific Conference Achievements in Mechanical and Materials Engineering, Nov. 29 - Dec. 2, Gliwice - Zakopane, pp.333-336.
• Humphreys, F. J., 1977, The nucleation of recrystallization at second phase particles in deformed Alluminium, Acta Metallurgica, vol. 25, n. 11, pp. 1323-1344.
• Kenney, M. P., Courtois, J. A., Evans, R. D., Farrior, G. M., Kyonka, C. P., Koch, A. and Young, K. P., 1988, Semisolid metal casting and forging, in Metals Handbook, 9^th ed., vol.15, pp. 327-338.
• Robert, M. H. & Kirkwood, D. H., 1988, Alloy slurry formation by partial melting, Proceedings of the Conference on the Solidification of Metals, pp. 373-376.
• Young, K. P., Kyonka C. P., Courtois, J. A., 1983, Fine grained metal composition, US Patent C.I. C22F ź, n. 4.415.374, New York, Mar. 30.
• Zoqui, E. J., Robert, M. H., 1998, Structural modifications in rheocast Al-Cu alloys by heat treatment and implications on mechanical properties, Journal of Materials Processing Technology, vol. 78, pp. 198-203.

Publication Dates

• Publication in this collection
  18 Mar 2004
• Date of issue
  Apr 2003

History

• Accepted
  July 2002