Effect of indentation size on plastic deformation processes - DOC

Document Sample
Effect of indentation size on plastic deformation processes - DOC Powered By Docstoc
					Effect of indentation size on plastic deformation processes in an ultrafine-grained Al-3% Mg alloy Zs. Kovács1, , N. Q. Chinh1, J. Lendvai1, Z. Horita2 and T. G. Langdon3
1: Department of General Physics, Eötvös University, Budapest 1117 Budapest, Pázmány P. sétány 1/A., Hungary 2: Department of Materials Science and Engineering, Kyushu University Fukuoka 812-8581, Japan 3: Departments of Aerospace & Mechanical Engineering and Materials Science University of Southern California, Los Angeles, CA 90089-1453, USA Keywords: Equal channel angular pressing (ECAP); ultrafine-grained structure; indentation; Portevin-Le Châtelier effect; grain boundary sliding

Abstract An Al-3% Mg alloy was processed by Equal Channel Angular Pressing (ECAP) for 8 passes in route Bc thereby reducing the grain size of the annealed sample from ~400 m to ~300 nm. The as-pressed ultrafine-grained material was examined using depth-sensing indentation tests over different length scales with the overall objective of identifying the different deformation processes taking place in the bulk submicrometer crystalline material. At small indentation sizes with indentation depths of <1000 nm, plastic instabilities in the form of the Portevin-Le Châtelier (PLC) effect were observed in the as-pressed material. At the macroscopic scale, however, the alloy exhibited no PLC instabilities thereby suggesting that dislocation motion then makes only a limited contribution to the plastic deformation and the flow is controlled instead by another mechanism associated with the presence of a very large number of grain boundaries.

I. Introduction The Equal Channel Angular Pressing (ECAP) technique is widely used to produce bulk submicrometer and nano-grained material from various metals and alloys [1-4]. During this process, the material is pressed through two equal channels which intersect within a die at a selected angle that is generally close to, or equal to, 90C. The large strain imposed in ECAP through consecutive pressings of the same sample allows a modification of the grain structure through the evolution of bands of subgrains into arrays of reasonably equiaxed grains and, ultimately, to the development of an ultrafine-grained microstructure. The development of this fine grain structure depends upon the nature of the pressing route which relates to the rotation of the sample between consecutive passes through the die. Three main routes may be identified: routes A, B and C corresponding to the rotation of the samples by 0, 90 and 180between each consecutive pass, respectively, where route B is further divided into routes BA and BC corresponding to rotations by 90 in alternate directions and by 90 in the same direction, respectively [5]. Among these possible routes, the evidence suggests that route BC leads most readily to an equiaxed grain structure [6] and this route was therefore used in the present investigation. The primary advantages of ECAP processing are that the reduction in grain size gives significant increases in the strength and the toughness of the material and, in addition, it may increase the formability at high temperatures. The forming of complex parts is often achieved by superplastic deformation where superplasticity is a diffusion-controlled process occurring only at relatively high homologous

Corresponding author: Zs. Kovács, e-mail:, Fax: +36-1-3722811

temperatures (T > 0.5Tm, where Tm is the absolute melting temperature) and at fairly low strain rates (typically of the order of ~10-4-10-3 s-1). Recently, it has been shown that the ultrafinegrained structures produced by ECAP provide an opportunity to achieve a superplastic formability at relatively lower temperatures (T of the order of ~0.5Tm) and at unusually rapid strain rates (d/dt > 10-2 s-1) in some alloys based on Al, Cu and Mg matrices [7-9]. In fine grained materials, superplasticity occurs primarily through the process of grain boundary sliding in which adjacent grains are displaced relative to each other and high strains are achieved without any significant grain elongation [10]. For polycrystalline materials, groups of grains may move in a cooperative manner producing cooperative grain boundary sliding (CGBS) through the collective motion of, typically, of the order of ~30 grains [11,12]. Aluminum-magnesium alloys are often used for studying the Portevin-Le Châtelier (PLC) effect and the associated plastic instabilities evident during tensile flow. The PLC instability occurs in solid solutions because of the solute-dislocation interaction [13]. In addition, the diffusion of solute atoms towards temporarily-pinned dislocations may increase the waiting time leading to the phenomenon of dynamic strain aging [14] and thus a negative strain rate sensitivity which is the physical basis of the plastic instability [15]. Macroscopically, the PLC effect is characterized by the repeated appearance of high strain rate deformation regimes (or bands) which require highly collective dislocation motion [16]. In load-controlled measurements, and as a consequence of the high strain rate, these bands appear in the load-displacement curve [17]. Very regular instability steps have been observed during depth sensing micro-indentation of a coarse-grained Al-3% Mg alloy [18] thereby permitting the detection of PLC instabilities down to very small plastic volumes. Earlier results using constant loading rate indentation showed that during measurements the PLC instability generally sets in at a constant loading time of 8 s in Al-3% Mg [19] and it can be observed over a wide range (up to about 2200 s loading time) which is roughly independent of the loading rate and of the indentation size [20]. Generally, the role of the loading time is crucial in indentation experiments with a constant loading rate with self-similar indenters (such as Vickers, Berkovich, conical) because the strain rate distribution in the inhomogeneously deformed plastic volume under the indenter is proportional to 1/t where t is the time [20]. In a simpler model, the equivalent strain rate can be defined as (d/dt)eq = (K/h)(dh/dt), where h is the indentation depth and K is a constant [21]. In constant loading rate experiments, if the hardness is constant the equivalent strain rate (d/dt)eq is 1/(2t) and this shows the importance of the loading time during the indentation.

II. Experimental A high purity (99.99%) Al-3% Mg alloy was processed using Equal Channel Angular Pressing (ECAP) for 8 passes in route Bc where the sample is rotated in the same direction by 90 between each pass [5]. The ECAP processing was conducted using a solid die with an internal angle between the two parts of the channel of 90. Using this internal angle, it can be shown that the imposed plastic strain is on each separate pass [22] leading to a total imposed strain of ~8. The initial grain size of the material was ~400 m before pressing but measurements after ECAP revealed an as-pressed grain size of ~300 nm. Further details of the ECAP process are given in a recent review [23]. The stability of the ultrafine-grained microstructure was investigated by conducting isochronal annealing treatments to the as-pressed samples for periods of 10 minutes from 60C up to 300C in incremental steps of 20C. Depth-sensing Vickers indentation experiments were made by using a Shimadzu DUH-202 ultra-microindenter. Different maximum loads (100 mN and 1000 mN) were applied to investigate the material over different length scales. The depth sensing indentation tests were carried out at constant loading rates chosen as 7 mN/s for the larger maximum load and 0.69 mN/s and 0.13 mN/s for the smaller maximum load.

III. Experimental Results and Discussion a. Large indentations Figure 1 shows the load-depth (F-h) curves obtained on an unpressed sample and on a sample after ECAP using an applied loading rate of 7 mN/s. In order to observe the strain rate fluctuations, which are a consequence of the PLC plastic instabilities, the indentation rate (dh/dt) is also shown in Fig. 1. For the annealed sample in Fig. 1(a), the regular steps appear in the depth-load curve as a consequence of the PLC plastic instabilities. Each instability step corresponds to a high indentation rate (dh/dt) peak while between the peaks the indentation rate is close to zero. The loading time when the PLC instabilities set in (the critical time for this condition) is about 8 s, and this value is in agreement with earlier results on coarse grained samples of the same alloy [19]. The corresponding indent size is then about 1-1.5 m. In the case of the sample after ECAP as shown in Fig. 1(b), the indentation curve is remarkably smooth up to the maximum load and the small fluctuations on the indentation rate curve originate from noise in the system. Thus, it is concluded that the ECAP sample does not exhibit PLC instabilities at least under these experimental conditions. Since the occurrence of PLC plastic instabilities requires dislocation motion and an appropriate interaction between the solute atoms and dislocations within the grains, the lack of any PLC plastic instabilities in the as-pressed sample leads to two possible conclusions: either the deformation is taking place through a flow process other than dislocation motion within the grains or the solute atoms have become depleted from within the matrix as a consequence of solute segregation to the very large volume fraction of grain boundaries that are present in this ultrafinegrained as-pressed alloy.



Fig. 1: Indentation depth and indentation rate as a function of load at 7mN/s loading rate in (a) the annealed sample and (b) the ECAP sample.

b. Small indentations Figure 2 shows similar indentation curves obtained on other ECAP samples in indentation tests conducted using much smaller maximum penetrations. In Fig. 2(a), the loading time regime is the same as in Fig. 1 to allow a direct comparison between the indentations with different sizes. Thus, the steps in the load-depth curve and the corresponding peaks in the load-indentation rate curve indicate the appearance of the PLC plastic instability at these small indentation depths. With increasing indentation depth, and at a depth of the order of ~1.5 m, the instability steps tend to disappear from the indentation curves. This effect is especially visible in Fig. 2(b) where the time scale extends to >700 s. It is also evident from these plots that the instability steps become more pronounced as the loading rate is decreased.



Fig. 2: Depth and the indentation rate as a function of load in ECAP Al-3% Mg (8 passes, route Bc) at ~1 m indentation size. The applied loading rates were (a) dF/dt = 0.69 mN/s, (b) dF/dt = 0.13 mN/s (Fmax=100 mN).

The appearance of the PLC plastic instabilities indicates that intragranular dislocation motion becomes important at sub-micrometer indent sizes. If there are instabilities at these small sizes, it suggests that the depletion of Mg atoms from within the matrix cannot be responsible for the disappearance of the PLC plastic instabilities at the large indentations. Therefore, it is reasonable to conclude that at large indentation sizes dislocation motion within the grains makes only a limited contribution to the plastic deformation and some other flow mechanism becomes rate-controlling where this other flow process must be associated with the ultrafine grain size and the presence of a large volume density of grain boundaries. Nevertheless, a depletion of Mg atoms from the matrix probably takes place in the material with a sub-micrometer grain size. This leads to the relatively small strain rate peaks during the appearance of the instabilities in the as-pressed samples in Fig 2. The results suggest that a transition from dislocation motion to some form of grain boundary process takes place at a penetration of ~1.5 m in depth as the plastic volume of the indentation increases. Thus, below and up to ~1.5 m of the indent size there is primarily dislocation motion within the grains but grain boundary processes become rate-controlling above this indent size Using the indent size related to the transition, it is possible to estimate the number of grains necessary for grain boundary processes to dominate through a simple calculation. The contact area, A, of the Vickers indenter is given by A  26.42 h2. A lower limit for the number of grains involved in the deformation may be calculated as N = 4 (A/d2, where d is the average grain size. For h = 1.5 m and d = 0.3 m, the value of N is ~800-900 grains, thereby indicating that in the as-pressed sample the deformation volume must extend to approximately 1000 grains in order that a grain boundary process becomes the dominant deformation mechanism. c. Effect of increasing the grain size through annealing An isochronal annealing treatment was applied to the ECAP samples to investigate the stability of the ultrafine-grained microstructure. Up to a temperature of 200C the PLC steps show the same behavior as in the as-pressed material. However, at higher temperatures the PLC effect becomes evident even for macroscopic indentation depths, as shown in Fig. 3 where tests are recorded after annealing at different temperatures from 180C to 300C. These results show the effect of the extensive grain growth that begins to occur at annealing temperatures close to 200C.

Fig. 3: Indentation rate obtained on the ECAP samples after isochronal heat treatments at 180C, 220C,260C and 300C ( the curves are shifted by 0.5 m/s subsequently).

Finally, it is necessary to address the question of the precise flow mechanism that occurs in these ultrafine-grained materials when the PLC instabilities are absent. Although the flow mechanism must be associated with the large volume of grain boundaries present in these materials, nevertheless conventional grain boundary sliding is controlled by the rate of diffusion and therefore this process cannot occur to any significant level during indentation testing of the Al-3% Mg alloy at room temperature. Furthermore, it has been shown in a series of extensive experiments on various Al-Mg and Al-Sc alloys that true grain boundary sliding with a strain rate sensitivity of ~0.5 occurs in only a limited number of alloy systems [24,25]. Nevertheless, there is evidence for the occurrence of relatively high tensile elongations in some Al alloys even when careful experiments show that conventional grain boundary sliding is not the rate-controlling process. For example, elongations of up to ~600% were recorded in some Al-Mg-Sc alloys when the measured strain rate sensitivity was only ~0.3 which is consistent with a dislocation glide mechanism [24]. These latter results are consistent with the present data and they show that some additional grain boundary process must occur in these ultrafine-grained materials in the as-pressed condition. More work is now needed to more fully elucidate the precise nature of this grain boundary mechanism. IV. Summary and Conclusions The effect of ECAP on an Al-3% Mg alloy was investigated by using a depth sensing Vickers indentation technique at room temperature. A transition was observed with increasing indentation size from dislocation motion at small indentations to a predominantly grain boundary effect at large indentations. The indentation depth approximating to the transition between these two mechanisms permits an estimate of the lower limit for the number of grains which must be involved in the deformation in order that a grain boundary process becomes rate-controlling. Applying isochronal heat treatment to the as-pressed material, the appearance of the PLC effect was observed even at large indentation sizes as a consequence of the onset of extensive grain growth at temperatures in the vicinity of 200C.

Acknowledgements The research of ZsK, NQC and JL was supported by the Hungarian Ministry of Education under contract number FKFP-0177/1999 and by the Hungarian National Scientific Research Funds under contract numbers OTKA-029644 and O29701.

References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] R. Z. Valiev, R. K. Islamgaliev and I. V. Alexandrov: Prog. Mater. Sci. Vol. 45 (2000), p. 103 M. Furukawa, Y. Iwahashi, Z. Horita, M. Nemoto and T.G. Langdon: Mater. Sci. Eng. A Vol. A257 (1998), p. 328 Y. Iwahashi, Z. Horita, M. Nemoto and T.G. Langdon: Acta Mater. Vol. 45 (1997), p. 4733 Y. Iwahashi, Z. Horita, M. Nemoto and T.G. Langdon: Acta Mater. Vol. 46 (1998), p. 3317 M. Furukawa, Z. Horita, M. Nemoto and T.G. Langdon: Mater. Sci. Eng. A Vol. A324 (2002), p. 82 K. Oh-ishi, Z. Horita, M. Furukawa, M. Nemoto and T. G. Langdon: Metall. Mater. Trans. A Vol. 29A (1998), p. 2011 M. Mabuchi, K. Ameyama, H. Iwasaki and K. Higashi: Acta Mater. Vol. 47 (1999), p. 2047 Z. Horita, M. Furukawa, M. Nemoto, A.J. Barnes and T.G. Langdon: Acta Mater. Vol. 48 (2000), p. 3633 K. Neishi, T. Uchida, A. Yamauchi, Z. Horita and T. G. Langdon: Mater. Sci. Eng. A Vol. A307 (2001), p. 23 T. G. Langdon: Mater. Sci. Eng. A Vol. A174 (1994), p. 225 M. G. Zelin and A. K. Mukherjee: Acta Metall. Mater. Vol. 43 (1995), p. 2359 O. A. Kaibyshev: Mater. Sci. Eng. A Vol. 324 (2002), p. 96 A. van den Beukel: Phys. Stat. Sol. (a) Vol. 30 (1975), p. 197 L. P. Kubin, K. Chihab and Y Estrin: Acta Metall. Vol. 36 (1988), p. 2707 P. Penning: Acta Metall. Vol. 20 (1972), p. 1169 P. Hähner: Mater. Sci. Eng. A Vol. A207 (1996), p. 208 Zs. Kovács, G. Vörös and J. Lendvai: Mater. Sci. Eng. A Vol. A279 (2000), p. 179 G. Bérces, N. Q. Chinh, A Juhász and J. Lendvai: J. Mater. Res. Vol. 13 (1998), p. 1411 G. Bérces, N. Q. Chinh, A Juhász and J. Lendvai: Acta Metall. Vol. 46 (1998), p. 2029 Zs. Kovács: PhD thesis, Budapest, Hungary (2002) M. S. Bobji and S. K. Biswas: Proceedings of the International Conference on Recent Advances in Metallurgical Processes, New Delhi (1997), p. 1223 Y. Iwahashi, J. Wang, Z. Horita, M. Nemoto and T. G. Langdon: Scripta Mater. Vol. 35 (1996), p. 143 M. Furukawa, Z. Horita, M. Nemoto and T.G. Langdon: J. Mater. Sci. Vol. 36 (2001), p. 2835 M. Furukawa, A. Utsunomiya, K. Matsugara, Z. Horita and T. G. Langdon: Acta Mater. Vol. 49 (2001), p. 3829 S. Lee, A. Utsunomiya, H. Akamatsu, K. Neishi, M. Furukawa, Z. Horita and T.G. Langdon: Acta Mater. Vol. 50 (2002), p. 553

Shared By: