Influence of Alfa Martensite on Shape Memory in Fe-Mn-Based Alloys

Transcription

1 Influence of Alfa Martensite on Shape Memory in Fe-Mn-Based Alloys Y. Tomota, K. Yamaguchi To cite this version: Y. Tomota, K. Yamaguchi. Influence of Alfa Martensite on Shape Memory in Fe-Mn-Based Alloys. Journal de Physique IV Colloque, 1995, 05 (C8), pp.c8-421-c < /jp4: >. <jpa > HAL Id: jpa Submitted on 1 Jan 1995 HAL is a multi-disciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

2 JOURNAL DE PHYSIQUE IV Colloque C8, suppl6ment au Journal de Physique 111, Volume 5, decembre 1995 Influence of Alfa' Martensite on Shape Memory in Fe-Mn-Based Alloys Y. Tomota and K. Yarnaguchi Department of Materials Science, Faculty of Engineering, Zbaraki University, Nakanarusawa-cho, Hitachi 316, Japan Abstract. An Fe-16MndSi alloy exhibits a good one-way and a little two-way shape memories by the stress induced y + E martensitic transformation and its reversion on heating and cyclic thermal y 2 E transformations, respectively. An intrusion of small amount of a' martensite at the initial shape change is not found to deteriorate the one-way memory and rather enhance the two-way memory. The two Fe-Mn-C alloys are also examined and their inferior shape memory is considered not due to the influence of a' martensite. 1. INTRODUCTION The shape memory in Fe-Mn based alloys is caused by the stress induced austenite(y) to epsilon martensite(&) transfomation and the E -+ y reverse transformation on heating. For obtaining a good one-way shape memory, following two are important conditions. (1) A large amount of the E martensite must be stress-induced to produce a shape change without an intrusion of usual slip. (2) The reverse transformation must take place by the backward motion of the Shockley partial dislocations which moved at the forward y + E transformation. Sato et al have studied the shape memory phenomenon first in an Fe-Cr-Ni single crystal[l]. The intrusion of a' martensite has been considered to result in imperfect memory. Then they changed the alloy to an Fe-30Mn-1Si single crystal to avoid the formation of a' martensite and discovered a perfect memoryf21. This finding has been developed to polycrystalline Fe-Mn-Si[3] and subsequently Fe-Mn-Cr-Ni-Si shape memory alloys[4] which are now commercially available. However, the roles of Si addition and so called "the training" treatment[5] given to the alloys are still unclear. Recently, Natsume et a1 have examined the effect of carbon addition to Fe-Mn and Fe-Mn-Si alloys[6]. It has been found that the degree of shape memory is improved in the Fe-Mn-Si-C alloy, but that in the Fe- Mn-C alloys is still poor. The reasons of the inferior memory were discussed from the view points of the intrusion of a' martensite and the strength of y against permanent slip. Quite recently, Tsuzaki et a1 have successfully summarized the influence of the solid-solution hardening on the shape memory based on the strength of austenite, where the hardening due to interstitial carbon atoms in the Fe-Mn- C alloys is not enough large in comparison with the strength of the conventional Fe-Mn-Si or Fe-Mn- Ni-Cr-Si alloys[7]. The shape change caused by the a' formation is of course irreversible and hence it is reasonable to conclude the intrusion of a' martensite lowers the degree of shape memory. However, it is still questionable whether or not a small amount of a' martensite inhibits the E -, y reversion. In fact, such a question remains in the understanding of poor shape memory in the Fe-Mn- C alloys. The purpose of this report is therefore to show how the a' martensite formed mainly at the intersections of E plates affect the shape memory associated with the y - E transformation. Article published online by EDP Sciences and available at

3 (3-422 JOURNAL DE PHYSIQUE IV 2. EXPERIMENTAL PROCEDURE The alloys used in the investigation were prepared by induction melting in an argon gas atmosphere. The results of chemical analysis of three alloys, 16Mn-6Si, 17Mn-0.2C, and 17Mn-0.3C, are shown in Table 1. The ingots were hot-forged and cold-rolled to 1.5mm thick sheets from which tensile specimens with the gauge section of 20 x 4 x 1.5mm were spark machined. The flow charts of heat treatments and testing procedure are illustrated in Fig. 1. The transformation temperatures of Ms and As are listed in Table 1. As shown in Fig.l(a), after austenitized at 1273K for 600s on a gear driven type tensile machine with an infrared-ray heating furnace, tensile specimens of 16Mn-6Si alloy were cooled down to 523K, at which they were extended by about 2% and then cooled to room temperature. During the cooling from 523Kboint B to C in Fig. l(a)), the applied load at 523K was held, so that the stress induced E martensite was yielded with elongation, i.e. initial shape change ALO. By this testing technique, the permanent slip can hardly take place during the transformation because the applied stress was below the yield strength. A few percent of plastic deformation of y at 523K is necessary to obtain a large amount of stress-induced E martensite; probably this is similar to the training treatment[6]. A typical example of the elongation vs. temperature curve is presented in Fig. 2 where the hu) was measured using an extensometer. Some specimens were subsequently deformed slightly at room temperature by ALr to induce the a' martensite. Either y + E or y + E + a' specimens cut from the tensile specimens were heated on a dilatometer. Similar heat was given to 17Mn-0.2C specimens(see Fig. l(b)). Since the Ms temperature is slightly higher than room temperature, 17Mn-0.3C specimens were deformed at room temperature after the austenitization (Fig. l(c)). In this alloy, the a' martensite is easily induced at the intersections of E plates. Their reversion treatment was the same with that used for 16Mn-6Si. The transformation behaviour with Table 1 Chemical compositions of the alloy used (mass%) and transformation temperatures(k) Austenitization 1273K 600s Austenitization 1273K 600s Austenitization 1273K 7.2ks C E-'Y Straining (y + E. a') a R.T. R.T. A W.Q. B (a) 0) (c) Fig. 1 Heat treatments and testing procedures: (a) for 17Mn-0.2C, and (c) 17Mn-0.3C.

4 cyclic warming and cooling was also examined. An example of the dilatometry result is shown in Fig. 3. The recovery in the length of a specimen deformed by the stress-induced transformation on heating(al1) was determined from this kind of data. When the two-way shape memory was observed, a repeated change in the specimen's length was measured as AL2. Here, the degree of one-way shape memory(dsr(1)) was calculated by either of following equations, DSR(1) = 100 x ALlIAI-0 (%) for y + E samples (1-4 DSR(1) = 100 x ALl/(ALO+hLr) (%) for y + E + a' samples (1-b) For discussing the two-way shape memory, a parameter DSR(2) is used. DSR(2) = 100 x AL2IALl (%I (2) The microstructure was observed by light, scanning (SEM), or transmission electron microscopy. For TEM observation, specimens were thinned by the twin-jet polishing technique. The foils were examined with a JEM 2000FX I1 microscope operated at 200kV. h ~ e l a l ~ t s l t ~ l,, l Temperature (K) Fig. 2 Change in the gauge length of a tensile specimen with an applied stress in the 16Mn-6Si alloy Temperature (K) Fig. 3 Change in the gauge length of a dilatometry specimen with heating and cooling in the 16Mn-6Si alloy having y+& microstructure. 3. RESULTS AND DISCUSSION 3.1 Microstructure The microstructure of as-quenched specimens was y + E in 16Mn-6Si and 17Mn-0.2C or nearly single y in 17Mn-0.3C. The a' martensite was easily yielded by the deformation at room temperature. Therefore, the special heat camer explained in Fig. 1 is necessary to obtain a large shape change due to E formation. Examples of a' martensite formed in an E + y microstructure are shown in Fig. 4. The a' martensites are observed scattering within E martensite plates. An a' grain recognized as dark area in Fig. 4(b) is confirmed by TEM observation(fig.4(a)). The volume fraction of a' martensite was approximately 4% which was determined by means of magnetic measurement technique, i.e., using a ferrite-scope. 3.2 Influence of a' martensite on the one-way shape memory Figure 2 shows the change in the length of a tensile specimen during the cooling under the applied stress in 16Mn-6Si. The length is first decreased due to thermal contraction effect and then increased due to the E formation. The start temperature of the traisformation, Ms, is higher than the Ms of the thermally induced, i.e., spontaneous transformation. Similar behaviour can be observed in

5 C8-424 JOURNAL DE PHYSIQUE IV a) b) Fig.4 Examples of 7 + E + a ' microstructure in the 16Mn-6Si alloy: (a)tem and (b)sem observations. 17Mn-0.2C. The recovery in the shape is determined from Fig. 3(the gauge lengths of specimens are different in Figs. 2 and 3 from each other). From these measured data, DSR(1) was calculated. The results are shown in Fig. 5. As can be seen from the figure, the influence of a small amount of a' martensite is negligible both in Fe-Mn-Si and Fe-Mn-C alloys. Even when the intrusion of a' martensite is avoided in Fe-Mn-C alloys, the shape memory is not enhanced. Consequently, their poor shape memory is not ascribed to a' martensite introduction but presumably to lower strength of the y phase. To obtain a good shape memory, the reversible motion of partial dislocations at-y - E transformation is required. Tomota et a1 have proposed the origin of this reversible movement is the back stress caused by the misfit strain[8]. If the y strength was low, the back stress would be possibly relaxed by the accommodation of permanent slip in front of an E plate tip. Then the reversibility of the partial dislocations would be lost and hence the DSR(1) would decrease. Tsuzaki et a1 have demonstrated that the solution-hardening in the y matrix governs the DSR or the recoverable strain[7]; the DSR is enhanced with an increase in the y strength. Another suspectable cause for the reversible motion is the short range ordering of Si atoms suggested by Sade et d[9] and now under investigation. The result of Fig. 5 implies that the intrusion of a' rnartensite does not become strong barriers against the backward motion of partial dislocations. It might be imagined that the dislocations move back remaining Orowan loops around a' particles. The influence of a' martensite is of course dependent upon the volume fraction. If the a' volume fraction was increased, for instance, up to more than lo%, the DSR(1) would be clearly decreased. a) b) Fig. 5 Influence of a' martensite on the degree of the one-way shape memory, DSR(l), as a function of initially given strain in 16Mn-6Si(a) and 17Mn-0.2C and 17Mn-0.3C(b).

6 3.3 Effect of a' martensite on the two-way shape memory The two-way shape memory has been found in Fe-Mn-Si alloys although the reversible strain is small[l0,ll]. An example observed in this investigation is presented in Fig. 3. In a particular temperature region, a specimen shows the cyclic shape memory only by warming and cooling without applied stress. If a specimen is heated up to a more elevated temperature, for example, 873K in Fig. 3, the two-way memory disappears so that the shortening of the sample is observed during the cooling. This shrinkage is corresponding to the volume change by the y + E transformation, that is, the essential dilatometry behaviour associated with the thermal transformation. It should be noticed that the Ms temperature for y - E cyclic transformation is higher than that in the solutionized sample. Moreover, the former Ms temperature is almost identical with that for the stress-induced y + E transformation observed in Fig. 2. When a specimen is once heated up to 873 K, the Ms temperature lowers as can be observed in Fig. 3. Here, two possibilities might be pointed out as the reason why the Ms temperature is higher in the case of the two-way shape memory. One is concerning the preferential nucleation of E martensite with a special variant. Since the density of one kind of partial dislocations with the same Burgers vector must become high reflecting the reversion from the stress induced E martensite[l2], the E plates with a preferential variant could be formed easily even by cooling with the absence of applied stress. The preferential shear at the Y-+E transformation results in showing elongation in a certain direction inspite of the volume shrinkage. This is, however, not easy to explain the fact that the Ms temperature is higher in the hysteresis loop of the two-way memory. Another possibility is the change in the back stress during the y - E transformation. If the internal stress generated around a tip of an E plate is relaxed to some degree, the opposite sense of shear at its reversion would cancel the pre-existed back stress and would furthermore produce the opposite sense of back stress. Then, this newly established back stress could aid the next y -+ E transformation, that is, the transformation is assisted by the internal stress. In this case, the back stress becomes zero at one moment during the transformation and its maximum decreases by the partial relaxation. This cyclic back stress formation model can well explain the twoway memory and the higher Ms although the experimental evidence is insufficient. On the contrary, the two-way memory is extremely poor in Fe-17Mn-0.2C, in which the back stress is not built up from the beginning because of easy operation of plastic relaxation due to the lower strength of y. A large part of the reverse transformation is considered to undergo by motion of the equivalent partial dislocations with different Burgers vectors on an identical (111) y, that is, (0001) E plane. a, e ~ 2 ~ 4 ~ 6 l 8 ~ ' ~ l " l 8 Tensile Strain (%) n Fig. 6 Effect of the intrusion of a' martensite on the degree of the two-way shape memory, DSR(2), in the 16Mn- 6Si alloy

7 JOURNAL DE PHYSIQUE IV The effect of the intrusion of a' martensite on the two-way memory is shown in Fig. 6. To be noted here is that DSR(2) is slightly increased by the existence of a' martensite. The cyclic building of the back stress must be enhanced by dispersed a' grains. 4. CONCLUSIONS The influence of small amount of a' martensite on one-way and two-way shape memory in Fe-Mn based alloys were examined and following conclusions have been obtained. (1) The effect of a' martensite intrusion on the one-way memory is of negligible order as long as the volume fraction is small such as 4%. The reversible motion of partial dislocation at y -;y E transformation is important for obtaining a good shape memory and the a' martensite grains do'not seem to become strong barriers against the reverse transformation. (2) A small volume hction of a' martensite intrusion is found to enhance slightly the degree of twoway shape memory. (3) The poor shape memory in Fe-Mn-C alloys in comparison with Fe-Mn-Si ones is not believed due to the intrusion of a' martensite but the y strength related with the reversible motion of the Shockley partial dislocations. Acknowledgments The authors appreciate Kobe Steel Co. for offerring the alloys used in this experiment. They wish to thank Prof. K.Tsuzaki of Kyoto University and Dr. S.Kajiwara and Mr. T.Kikuchi of National Research Institute for Metals(Japan) for helpful discussions. They are also grateful to Dr. T.Noguchi for help in TEM observations at the Center for Instrumental Analysis, Ibaraki University. References [I] A.Sato, H.Kasuga and T.Mori: Proc.ICOMAT 1979,Cambridge, MA, U.S.A., pp [2] A.Sato, E.Chishima, K.Soma and T.Mori: Acta Metall. 30(1982) [3] M.Murakami, H.Otsuka, H.G.Suzuki and S.Matsuda: Proc.ICOMT 1987, pp [4] Y.Moriya, T.Sampei and LKozasu: Abstract of Spring Meeting of JIM, 1989, p.222 [5] H.Otsuka, M.Murakami and S.Matsuda: Proc. Int. Meeting on Advanced Materials, Tokyo, Materials Research Society, 9(1989) pp [6] Y.Natsume, K.Tsuzaki, and T.Maki: Abstract of Spring Meeting of JIM, 1992, p.218 [7] K.Tsuzaki, Y.Natsume, Y.Tomota and T.Maki: accepted for publication in Scripta Metall. Mater. [8] Y.Tomota, M.Piao, T.Hasunuma and Y.Kimura: Jour.JIM, 54(1990) [9] M.Sade, K.Halter and E.Hombogen: Z.Metallk. 79(1988) [lo] Y.Tomota, W.Nakagawara, K.Tsuzaki and T.Maki: Scripta Metall.Mater. 26(1992) [ll] T-Shiming, LJinhai and Y.Shiwei: Scripta.Metall.Mater., 25(1991) [12] A.Sato: Proc.Int.Meeting on Advanced Materials, Tokyo, Materials Researc Society, 9, Shape Memory Materials (1989)pp