Int J Mater Form (2009) Vol. 2 Suppl 1:765–768 DOI 10.1007/s12289-009-0597-3 © Springer/ESAFORM 2009

Influence of the chemical composition on the Thixoformability of steels J.C. Pierret1, A. Rassili1, G. Vaneetveld1, J. Lecomte-Beckers2, G. Walmag3 1

ThixoUnit ULg, PIMW B56, University of Liège - Sart Tilman 4000 Liège Belgium 2 MMS ULg, B52, University of Liège - Sart Tilman 4000 Liège Belgium 3 CRM, PIMW B56, Sart Tilman 4000 Liège Belgium

ABSTRACT: This work deals with the qualification of a variety of steels for their shaping by the thixoforging process. This technology requires setting up a globular microstructure inside the material during reheating to a temperature between the solidus and liquidus. As the evolution of the liquid fraction is strongly connected to the steel composition, it is useful to understand how low carbon steel could be alloyed and still thixoformable. The most critical parameter for this is the carbon content. In this study, a theoretical analysis of the phase evolution during the reheating has been performed on the MT Data software to investigate the influence of alloying elements. These first results have been confirmed by Differential Thermal Analysis and by inductive heating experiments on steel slugs. Finally, some parts have been shaped using a thixoforming tool mounted on a hydraulic press. Micrographs of reheated slugs as well as of actual parts are also presented in this paper. KEYWORDS: Thixoforming, thixoforging, steel.

1 INTRODUCTION Thixoforming is the shaping of metal components in the semi-solid state. This forming technology requires a globular microstructure in the material in order to achieve the shear thinning behaviour. In this work, we compare the thixoformability of two classical steels and their modified grades (C38 and 100Cr6) and of a low carbon steel, the SAE1008 one. Table 1 gives the chemical composition of this steel.

Even if these elements have an interesting impact on the semi-solid range, they are not always beneficial. On figure 1, for instance, we see that, at low liquid fraction, the curve is steeper when the alloying content is higher. As we are working at low liquid fraction (between 15 and 20 %) we could prefer the low alloyed version. 1,20E+02

1,00E+02

Table 1 : Chemical composition of SAE1008 steel

C 0,068 Cr 0,067

Si 0,222 Ni 0,024

Mn 0,427 Al 0,020

P 0,016 Cu 0,050

S 0,006 Fe Balance

2 THEORICAL STUDY OF THE COMPOSITION Previous works have shown the influence of the chemical composition on the liquid fraction vs. temperature curve [1,2]. A similar study has been realised for the low carbon steel on the MTData software (available from National Physical Laboratory UK). Unfortunately, this software calculates only equilibrium results. For example, figure 1 shows the effect of an increase of the Si or Mn content to 1% on the liquid fraction. Table 2 summarizes the observed effect for five alloying elements (C, Mn, Si, Cr and P) when their contents increase. ____________________ * Corresponding author: postal address, phone, fax, email address

8,00E+01

Fraction liquide [%]

Element % Element %

Composition initiale 1% Si 1% Mn

6,00E+01

4,00E+01

2,00E+01

0,00E+00 1450

1460

1470

1480

1490

1500

1510

1520

1530

1540

-2,00E+01

Température [°C]

Figure 1: Effect of a modification of the steel composition

In the other hand, alloying elements could have an impact on the physical properties of the material. For instance, Si increases the yield stress and the oxidation resistance while Mn increases mechanical properties but also solidification shrinkage. Moreover these elements

1550

766 decrease thermal and electrical conductivity, which are important for inductive heating. Table 2 : Effect of the alloying elements (TS and TL are the solidus and liquidus temperatures, T50% is the temperature where solid and liquid contents are equals and df is the derivative of the liquid fraction/temperature curve)

Ele me nt C Mn Si Cr P

TS

T50%

TL

~

~ ~

(df/dT) 10%

(df/dT) 50%

T50 %TS

Figure 3: Liquid fraction vs. temperature curve obtained by DTA for the SAE1008 steel

However, a chemical modification of the alloy could be of first interest as the one made on C38 steel by Ascometal [3]. As shown on figure 2, this modification (C38 LTT) increases the semi-solid interval, decreases the working temperature and flatter the curve at low liquid fraction (until 50%).

3 DIFERENTIAL THERMAL ANALYSIS 1

4 INDUCTIVE HEATING In order to approach industrial requirements, we have heated slugs in an inductive furnace. The slug was 30mm diameter and 59mm height. Figure 4 shows the temperature evolution in the centre of the slug and on a point at mid-height and at 4 mm from the lateral surface. It could be seen that the temperature homogeneity is quite good at the end of the heating cycle, so we could expect a quite homogeneous structure. 1600

liquid fraction

0.8 0.6

1400

C38LTT

1200

C38

0.2 0 1380

Temperature [°C]

0.4

1000

800

600 Surface / mid-height Centre 400

1430 1480 Temperature °C

1530

Figure 2: LTT steel vs. classical one

200

0 0

20

40

60

80

100

Time [seconds]

Figures 2 and 3 show the results obtained by Differential Thermal Analysis for C38, C38LTT and SAE1008 steels. Compared to the MTData calculations, these results are interesting because there are not equilibrium ones. But, as the heating rate is only 10°C/minute, they are still not perfectly relevant for industrial applications for which fast heating is needed. Indeed, heating rate has an impact on the material structure and behaviour [4]. When the semi-solid range of C38LTT is higher 100°C, it is of only 50°C for SAE10008. So the heating must be rather more accurate for low alloyed steels than for high or medium alloyed ones.

Figure 4: Heating cycle for the SAE1008 steel

Figure 5 shows the microstructures observed in the quenched slug. Must of the material has a structure close to the one on fig. 5 b but on different areas, the structure is the one of fig.5 c or d, so a globular structure but where primal rolling structure is still visible. Also, in the centre of the slug some areas have still a totally laminar structure, even if some liquid is visible on grains boundaries (fig. 5 a). For this steel, we haven’t managed to obtain a full globular microstructure, which was possible for the classical and LTT versions of C38 and 100Cr6 steels.

120

767

a

b

c

d

Figure 5: Structure in SAE1008 steel in the centre of the slug (a), at mid-height and mid-radius (b), at mid-height and close to the vertical axis (c) and on the axis and close to the top (d)

5 Parts forming

Figure 7: Defects observed close to the surface

Béchet-Beaujard etching exhibits the liquid areas in black where solid grains appear lighter. For all the grades, we observe the classic globular structure allowing the appropriate shear thinning behaviour [7].

For this work, we used a tool developed in a previous research [5] to shape parts with the different steels. In these parts, we observed after polishing few inclusions for all the steels. We also found some pores in the parts in SAE1008 steel but even for this grade, only 3 or 4 porosities were found on the cutting plane (figure 6).

Figure 8: Grain sizes observed by Nital etching close to the surface (a) and in the centre (b) of the part

Figure 6: Inclusions and pores observed in the parts

On the polished half-parts, we noticed some oxide inclusions close to the surface and some surface defects as folds (figure 7). We could not establish any correlation between the steel grade and these defect occurrences. But for the SAE1008 steel, we also observed inverse extrusion around the forming punches, due to liquid segregation, on some parts when forming speed was high. Nital etching allows visualisation of the primary structure of the material. In each part, we observe globally homogeneous pearlite structure. Figure 8 shows the structures for the SAE1008 steel (the one for which the structure is the least homogeneous) close to the part’s surface and in the centre. Even for this grade, the grain sizes are of the same order of magnitude, the grains in the centre are around 2.5 times larger than the ones close to the surface. Moreover, the homogeneity of the structure could be easily improved by a suitable thermal treatment [6].

Figure 9 shows the SAE1008 steel in an area where the material was only compressed and in an area where the material flowed in a cylinder perpendicular to the compression axe. It could be seen that the liquid content is a bit more important in the outside of the part but the difference is not too important. The difference is still lower for the other grades, especially for the C38 and C38LTT steels, for which, with an appropriate choice of forming parameters, the structure is close to completely homogeneous, except close to the surface where thermal exchanges with the tool drive the structure. Due to the structure obtained after heating and especially the fact that in some areas the rolling bands were still presents (figure 4), we could observe the material flow thanks to the bands of deformation in the shaped parts for the SAE1008 steel. These bands have been observed only in the compressed area of the parts, where the material present in the centre of the slug lies after forming. These bands haven’t been observed for the other grades as their structure was more homogenous at the end of the heating step.

768

Figure 9: Globular structure revealed by BéchetBeaujard etching

The recorded forming loads are essentially the same for all the steel grades. It was a bit lower for the 100Cr6. The load is mainly driven by the liquid content and the forming parameters which were the same for all the grades. But it also depends on the thermal exchanges with the tool [8,9] which are higher if the material is hotter because the tool is kept to a constant temperature. This explains why the load was a bit higher for the C38 steel that must be heated up to a higher temperature. For the SAE1008 grade, the rolling bands, still present in the material at the end of the heating, increase the load needed to shape the part as this material can’t flow like a fully globular one. Nevertheless, the difference of load is only of 10 to 15 %, so the global behaviour is essentially the same.

Figure 10: Flow lines observed in SAE1008 parts

6 CONCLUSIONS It is well established that the steel composition has a great influence on its thixoformability. It is also clear that a highly alloyed material is usually easier to shape by a semi-solid process than a low alloyed one thanks to the larger semi-solid range. Nevertheless, we’ve showed that it is possible to apply the thixoforming technology to a very low alloyed grade even if the structural homogeneity is lower and the load needed slightly higher. This requires anyway a more accurate heating strategy.

ACKNOWLEDGEMENT The authors gratefully acknowledge the University of Liège, the COST 541 action, the First Europe Project and the Walloon Region for their financial support.

REFERENCES [1] C. Fraipont, J. Lecomte-Beckers, Determination of the Influence of Alloying Elements on Solidification Parameters Used for the Study of the Thixoformability of a Chromium Steel, Proc. of 10th ESAFORM Conf. on Material forming (2007) pp113-1117

[2] C. Fraipont, J. Lecomte-Beckers, Influence of Alloying Elements on the Thixoformability of a Chromium Steel, Proc. of 11th ESAFORM Conf. on Material forming (2008) [3] A.Rassili, M.Robelet, R.Bigot, D.Fischer, Thixoforming of Steels and Industrial Applications, Proc. of 10th ESAFORM Conf. on Material forming (2007) pp1136-1141. [4] W. Püttgen and W. Bleck : DTA-Measurements to Determine the Thixoformability of Steels, Steel Research Int. 75 (2004) No. 8/9 pp 531-536 [5] D. Fischer, B.A. Behrens, B. Haller, A Rassili, J.C. Pierret, H. Klemm, A. Studinski, B. Walkin, M. Karlsson, M. Robelet, L. Natale and F. Alpini : Series production of thixoformed steel parts, Trans. Tech Publications, Solid State Phenomena Vols. 116-117, Switzerland, (2006) pp 618-621. [6] JC Pierret, A. Rassili, G. Vaneetveld, P. Cézard, R. Bigot, Steel thixoforming, influence of forming parameters on the microstructure, Proc. of 11th ESAFORM Conf. on Material forming (2008) [7] P.A. Joly, R. Mehrabian : The rheology of a partially solid alloy, Journal of Materials Sciences 11, (1976), 1393-1418 [8] E. Becker, R. Bigot, L. Langlois, V. Favier, JC. Pierret, P. Cezard, Metallurgical and mechanical analysis from thixoforging steel shape, Trans. Tech Publications, Solid State Phenomena (2008), 701706 [9] JC. Pierret, A. Rassili, G. Vaneetveld, J. LecomteBeckers, R. Bigot, Robustness of a thixoforming production line, Trans. Tech Publications, Solid State Phenomena (2008), 207-212