Development of new routes of severe plastic deformation through cyclic expansion–extrusion process

Elsevier Editorial Systemtm for Materials Science & Engineering A Manuscript Draft Manuscript Number: Title: Development of new routes of severe plastic deformation through cyclic exp

See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/264049921

Development of new routes of severe plastic

deformation through cyclic expansion–extrusion process

Elsevier Editorial System(tm) for Materials Science & Engineering A

Manuscript Draft

Manuscript Number:

Title: Development of new routes of severe plastic deformation through cyclic expansion-extrusion process

Article Type: Research Paper

Keywords: Severe plastic deformation; Cyclic expansion extrusion; Aluminum; Mechanical properties; Microstructure; Micro shear band

Corresponding Author: Dr Ramin Ebrahimi,

Corresponding Author's Institution: Shiraz University

First Author: Nima Pardis, PhD Candidate

Order of Authors: Nima Pardis, PhD Candidate; Cai Chen, PhD Student; Mehrdad Shahbaz, PhD

Candidate; Ramin Ebrahimi; Laszlo Toth, Professor

Abstract: This paper introduces two new processing routes for a recently introduced severe plastic deformation technique, cyclic expansion extrusion (CEE) Two processing Routes (I and II) were experimentally performed on aluminum alloy 1050, the processed samples were investigated and compared in terms of their microstructural and mechanical characteristics A significant improvement

in mechanical properties was observed after one CEE pass via different processing routes Different grain structures were achieved after Routes I and II showing a more homogeneous microstructure and hardness distribution in Route II compared to Route I In addition, compression tests of the processed samples demonstrated that Route II results in a homogeneous compressive strength Finally,

microstructure evolution during subsequent passes of this process was investigated by electron back scattered diffraction Micro shear bands were found as potential sites for accelerating the formation of new grains which resulted in fragmentation of the initial grains and leading to an ultrafine-grained (UFG) microstructure

April 21 ,2014

Dear Editor

I am sending you our new paper entitled: “Development of new routes of severe plastic deformation through cyclic expansion-extrusion process ” to be considered for publication in Materials Science and Engineering: A The manuscript is the authors’ original work and has not been published in any journal

nor has it been simultaneously submitted elsewhere

Sincerely yours,

Dr Ramin Ebrahimi, Associate Professor, Department of Materials Science and Engineering, School of Engineering,

Shiraz University,

Cover Letter

Development of new routes of severe plastic deformation through

cyclic expansion-extrusion process

N Pardis*1, C Chen2,3, M Shahbaz1, R Ebrahimi†1, L.S Toth‡2,3

1 Department of Materials Science and Engineering, School of Engineering, Shiraz University, Shiraz, Iran

2 Laboratoire d'Etude des Microstructures et de Mécanique des Matériaux (LEM3), UMR 7239, CNRS / Université de Lorraine, F-57045 Metz, France

3 Laboratory of Excellence on Design of Alloy Metals for low-mAss Structures (DAMAS), Université de Lorraine,

France

Abstract

This paper introduces two new processing routes for a recently introduced severe plastic deformation technique, cyclic expansion extrusion (CEE) Two processing Routes (I and II) were experimentally performed on aluminum alloy 1050, the processed samples were investigated and compared in terms of their microstructural and mechanical characteristics A significant improvement in mechanical properties was observed after one CEE pass via different processing routes Different grain structures were achieved after Routes I and II showing a more homogeneous microstructure and hardness distribution in Route II compared to Route I In addition, compression tests of the processed samples demonstrated that Route II results in a homogeneous compressive strength Finally, microstructure evolution during subsequent passes

of this process was investigated by electron back scattered diffraction Micro shear bands were found as potential sites for accelerating the formation of new grains which resulted in

fragmentation of the initial grains and leading to an ultrafine-grained (UFG) microstructure

1 Introduction

Severe plastic deformation is considered as a powerful processing tool for producing bulk ultrafine grained (UFG) / nanostructured materials This approach is based on giant straining of bulk metallic materials [1-4] Most SPD techniques are classified as batch processing methods in which strain accumulation is achieved by performing consecutive passes From an industrial point of view, batch SPD techniques such as equal channel angular pressing (ECAP) [1,5-7], might seem less interesting compared to other continuous processes like high pressure torsion (HPT) [8] or high pressure tube twisting (HPTT) [9], where there is no need to repeat the process several times to reach the desired amount of accumulated strain On the other hand, batch techniques like ECAP may give an opportunity to define different processing routes between consecutive passes of the process which can be an effective tool in controlling and manipulating the resulting microstructure [5,6] Generally, these processing routes are simply performed by rotation of the sample around its main axis Such rotation is applicable in some SPD techniques like ECAP [5,6] and simple shear extrusion (SSE) [10,11], which do not have axisymmetric die geometry However, in other techniques with axisymmetric die geometry, like cyclic extrusion compression (CEC) [12], cyclic expansion extrusion (CEE) [13], or tube channel pressing (TCP) [14], the processing route would be limited to reversing the sample orientation/pressing direction with respect to die between consecutive passes [15] This paper presents a non- axisymmetric version of the CEE technique which makes it possible to introduce new processing routes of this technique for processing samples with rectangular cross section The resulting mechanical and microstructural evolutions are presented and discussed

2 Principles of different processing routes in CEE

Two major processing routes are defined for CEE processing of samples with rectangular cross section These routes are nominated as Route I and Route II which are illustrated in Fig 1

As can be seen in Fig.1, CEE processing in Route I is performed under plane strain conditions and therefore, both steps (expansion and extrusion) are performed in the same plane (Fig 1a) On the other hand, in Route II, expansion and extrusion steps take place on different planes which

are normal to each other (Fig 1b) Although each step (expansion or extrusion) is performed in plane strain condition, the overall process in Route II cannot be considered as a plane strain operation Based on Fig 1, a sample of a×b cross section is expanded in plane strain condition to

a square of b×b , which is subsequently extruded in plane strain condition to a rectangular cross section of a×b and b×a through processing Routes I and II, respectively Therefore, the amount

of von Mises accumulated strain in one pass can be calculated for both routes as:

This relation, however, is an average deformation value across the whole section in which shear components are neglected Namely, it is not excluded - and will be shown below - that the deformation can be heterogeneous leading to smaller strains in the center region and larger strains in the external part of the sample

3 Experimental procedure

3.1 Material and processing

Samples of 10 mm × 20 mm × 60 mm were machined from a 1050 aluminum alloy strip They were annealed at 600˚C for 2 h and furnace cooled to room temperature A split die configuration (consisting of two similar extrusion dies) was designed and used for CEE processing, suitable for both routes (Fig 2)

Using this die set up (Fig 2), a sample with 10 mm × 20 mm cross section is expanded in plane strain condition to a square with 20 mm side length following by an extrusion stage during which the sample regains its 10 mm × 20 mm initial section In this way, the process can be repeated several times According to Eq (1), the dimension values of the sample result in an average imposed strain value of for each pass of the process Route I was performed by placing the two die halves in plane strain configuration (Fig 1a), while 90 degree rotation of one half with respect to the other changes the configuration from Route I into Route II (Fig 1b)

After assembling the die halves in a desired configuration (Route I or II), the process was started

by pressing a "sacrifice" sample to fill the blocked expansion-extrusion chamber Subsequently, the exit channel was unblocked and consecutive pressing of other samples was performed More details of these processing sequences are described and illustrated in Ref [13] To reduce the

friction at the die-sample interface, each sample was wrapped with Teflon tape before inserting it into the die channel The process was performed with a pressing speed of 0.2 mm/s

3.2 Microstructural studies

Different planes on the processed samples were polished to a mirror like surface and etched with modified Poulton’s reagent [16] to reveal the corresponding microstructure using a stereo microscope The polished surfaces of samples were also electro-etched with Barker reagent [16] and subsequently investigated using polarized optical microscopy Electron back scatter diffraction (EBSD) technique was used for detailed examination of the microstructure The investigations were performed using JEOL 6500F scanning electron microscope (FEG-SEM) equipped with a field emission gun operating at 15 kV The steps of the sample preparation for EBSD were mechanical grinding (using 500-4000 grit size SiC papers) followed by polishing the surface with diamond compounds of 9 μm, 3μm and 1μm, subsequently The samples were further electro-polished in an electrolytic solution composed of 10 ml perchloric acid and 90 ml ethanol at 263K with a DC voltage of 18V to obtain a mirror-like surface The EBSD observations were accomplished with step size of 0.1 μm For processing the measurements, the HKL acquisition software was used

3.3 Mechanical studies

Vickers microhardness measurements were performed by applying 25 g load with a loading rate

of 5 g/s and 15 s dwell time The tests were carried out across the thickness of the samples on the ED-ND plane (Fig 1) with an incremental distance of 0.5 mm The indentation was repeated four times at each location and the average microhardness value was calculated Compression tests were performed on rectangular parallelepipeds of 10 mm × 10 mm × 15 mm extracted along the extrusion (ED) and transverse (TD) directions to give compression planes on TD-ND and ED-ND, respectively Fig 3 illustrates the orientations of the compression specimens with respect to a CEE processed sample The side length in the cross section of these parallelepipeds was considered equal to thickness of the CEE processed billet (i.e 10 mm) to maintain various grain structures across the thickness of compression sample Subsequently, the length of main

axis in these samples was chosen as 15mm to give the recommended aspect ratio of height/width equal to 1.5 [17]

4 Results and discussion

Fig 4 illustrates the in situ geometry of the samples for the two processing Routes I and II Geometrical and dimensional examination of these samples approved that for both processing routes, the expansion-extrusion chamber was satisfactorily filled and therefore, the designed amount of strain ( ) was accumulated in the samples in one pass The sample, however, does not recover its rectangular shape at the end of the deformation process because its end sections become U-shape; see in Fig 3 This shape is due to strain heterogeneity; there is more strain in the outer regions along the ND direction which leads to more stretching in the ED direction and produces the U-shape end in the samples

Consideration of the deformed geometry of an initially cubic elements (Fig 1) shows that in Route I, the material experiences a cyclic deformation path as each element of the material expands and extrudes in the same plane and by the same amount of strain (Fig 1a) However, the situation is different for Route II, as expansion and extrusion steps take place on different planes (Fig 1b) Therefore, material elements experience no cyclic deformation after one CEE pass via Route II This fact is illustrated in Fig 1 by considering distortion of cubic elements of a sample

at different stages of deformation for each processing route Therefore, processing Route II is expected to be more effective for grain refinement which will be verified below by microstructural investigations

The results of microstructural investigations obtained by optical microscopy are illustrated in Figs 5 and 6 Fig 5a shows a typical microstructure of the undeformed sample in TD-ND plane (the plane normal to the extrusion direction) The microstructures in the same plane after performing one pass in processing Routes I and II are illustrated in Figs 5b and 5c, respectively

It can be seen that the grain size and morphology after both processing routes are significantly different compared to the annealed structure For Route I, grains are equiaxed at the vicinity of the center line, while elongated grains are visible after processing Route II which are oriented

along TD (Figs 5c and 5d) This observation is in good agreement with thepredicted

configuration of initially cubic elements in ideal deformation condition after different processing routes (Fig 1) which only considers normal strain components However, a more severely

deformed structure is observed at regions adjacent to the surface of processed samples after both processing routes The corresponding microstructure viewed at ED-ND plane is illustrated in Fig

6

Similar to Fig 5b, it is seen that the through thickness structure after Route I is not uniform While equiaxed grain structure is observed at the central parts of sample after Route I, grains are severely elongated adjacent to the lateral boundaries of the die This excessive deformation is due to a redundant shear deformation which is dependent on the amount of reduction, semi die angle as well as the die profile [18]

The excessive shear deformation is more visible in the outer region after processing Route I compared to Route II (Figs 6b and 6c) The more uniform structure in Route II can be attributed

to the change in orientation of the extrusion plane with respect to the expansion plane in this route (Fig 1) As a result, the through thickness direction of the extruded sample would be the plane strain direction of its previous expansion step and therefore a more uniform structure is expected across the thickness In addition, the differences in orientation and aspect ratio of the expanded elements with respect to the extrusion direction for Routes I and II (Fig 1) might be another reason for such differences in microstructures

Although redundant strain would intensify the degree of severe plastic deformation, it would affect structural homogeneity and uniformity of the resulting mechanical properties across the thickness of samples This fact was investigated by analyzing the variation of microhardness values on ED-ND plane and the microstructures which are illustrated in Fig 7 for the two processing routes

It can be seen that Route II results in a more uniform hardness distribution across the thickness together with a more uniform grain structure on the ED-ND plane On the other hand, the hardness value is higher near the surface of the sample after Route I and gradually decreases toward the center Such hardness gradient is attributed to the microstructural inhomogeneities across the thickness resulting from high amount of shear strain near the surface Despite the differences in hardness homogeneity after these processing routes, the average hardness values after these processing routes are close to each other (53 HV and 53.7 HV for Route I and Route

II, respectively) and significantly higher than the average value in the annealed condition (29 HV)

Microstructural and microhardness investigations indicated that Route II leads to a more uniform structure after one pass This fact was further confirmed by performing compression tests at different orientations The compression test results are displayed in Fig 8 indicating the uniformity of compressive strength after processing Route II This might be due to the uniformity of deformation throughout the cross section as well as the similarity of grain orientations with respect to the different compression directions (ED and TD) (Figs 5c and 6c) After processing Route I, however, the grain structure on the planes normal to ED and TD are not similar (Figs 5b and 6b) Despite the equiaxed structure in the central part of this sample, the majority of grains are elongated in the extrusion direction at the vicinity of central band which is due to additional shear component of strain imposed by the die geometry at these areas Therefore, anisotropy in properties is seen in the compressive strength after processing Route I

It is also evident from Fig 8 that after one CEE pass the yield strength is significantly increased compared to its value in the annealed condition Such considerable increase in compressive strength as well as the significant microhardness improvement and structure refinement, even after one pass of CEE, indicate that the CEE process can be used as an effective SPD technique for producing bulk ultrafine grained materials

Due to the observation of a more homogeneous structure and properties after processing Route

II, more detailed microstructure investigations were performed on samples processed by this route (Route II) As an example, the microstructure of a sample after the first CEE pass via Route II was investigated by polarized light as well as electron backscatter diffraction (EBSD) and shown in Fig 9 These images were taken at the vicinity of the center line on the ED-ND plane after the first pass through Route II This position is schematically illustrated in Fig 9

Some bands are visible in both images along which crystallites of submicron size are evident These bands can be considered as deformation heterogeneities where dislocations form a quasi periodical structure consisting of geometrical necessary dislocation (GNDs) These GNDs are suspected to fall into walls and create in this way new grains [19] Crystallites of low angle

boundaries are expected to convert subsequently into ultrafine grains by gradual increase in their boundary misorientation angle during further passes of the process Indeed, one can see in Fig 9c that already after one pass in Route II there is a large proportion of high angle boundaries between the newly formed grains reaching a fraction of about 67% It should be noted that in this work the misorientation distribution was calculated between neighboring grains, not between neighboring pixels Such misorientation analysis was introduced by Toth et al [20] which has more physical meaning than the usually shown pixel-to-pixel misorientation distribution In this method, grains (with grain boundaries of at least 5° misorientation) are identified, then the misorientation between grains is calculated as the misorientation between the average orientations of the grains

Fig 10 shows Fig 9b in form of band contrast map Careful inspection of both figures reveals the presence of shear bands along which deformation localization can result in enhanced continuous dynamic recrystallization (CDRX) due to the larger local deformation Examples of such shear bands are highlighted in Figs 10b and b' This fact is in agreement with a model for the strain-induced formation of UFGs [21] which states the formation of new grains at micro shear bands (MSBs) at the initial stages of deformation It is important to note that the mechanism is not nucleation and growth of grains (discontinuous dynamic recrystallization, DDRX) but CDRX by forming new boundaries Subsequently, the fraction of high angle boundaries increases with increase in the density of MSBs at higher strain values, examples are illustrated in Fig 11 It is clear that increasing the amount of applied strain after the second CEE pass results in a higher density of MSBs as well as some intersecting MSBs (Fig 11) which is believed to help the generation of a more uniformly subgrain structure As a result, the UFG structure spreads through the whole volume [22] Such uniform distribution of UFG grains is illustrated in Fig 12 which shows EBSD micrograph after four CEE passes through Route II

Equiaxed grain structure with high angle boundaries and an average grain size of ~720 nm was achieved after four extrusion passes (Fig.12) Careful inspection of this image reveals some intersecting bands along which higher density UFGs are visible This result, together with the previous ones (Figs 9, 10 and 11) confirms the recommended model for strain-induced UFG

formation based on the evolution of MSBs [21] as well as occurrence of enhanced CDRX in these areas

The next-neighbor misorientation distribution of samples after four passes (Fig 13) shows a nearly constant frequency as a function of misorientation angle It does not mean, however, that the grain orientation distribution is random Similar result was also reported for processing of pure aluminum by HPTT [23] for which the accumulated strain was about the same as the strain value in this study As seen in Fig 13, there is very significant difference of the obtained distribution from the Mackenzie distribution which corresponds to the perfectly randomly oriented neighbor case There is a significant evolution of the misorientation distribution between the first and fourth passes (see Figs 9c and 13); in the fourth pass, there is more high angle boundaries: the more than 15° boundaries are now represent about 84% of the whole distribution A further observation from Fig 13 is that at the very large angles - near 60°- the distribution follows the Mackenzie one very precisely while at the lower angles, below 30°, the measured frequency is higher than the Mackenzie one The latter range of misorientation belongs

to grains that are within larger grains (entirely inside, not at the original grain boundary) meaning that at this stage there is still a grain fragmentation process going on, the steady state is not yet reached It has been shown by Pougis et al [24] in copper that in the steady state the misorientation distribution becomes very similar to the random Mackenzie

5- Summary and conclusions

New processing routes (I and II) were developed for severe plastic deformation of samples with rectangular cross section by cyclic expansion-extrusion (CEE) The samples were processed successfully and further studied by conducting mechanical and microstructural investigations Mechanical properties were significantly improved in both processing routes due to the considerable changes observed in grain structure and size Redundant shear deformation was more visible near the surface of the samples after processing Route I which resulted in non-uniformity of the mechanical properties However, a more homogeneous microstructure as well

as uniform compressive strength and hardness distribution were achieved in processing Route II Micro shear bands (MSBs) were found as regions where the formation of new grains is accelerated by CDRX within the original grains Further CEE processing introduced a high

density of intersecting MSBs in the grain interiors and resulted in fragmentation of the original grains to ultrafine grains with high angle misorentation boundaries

Acknowledgements

This work was supported by the French State through the program "Investment in the future" operated by the National Research Agency (ANR) and referenced by ANR-11-LABX-0008-01 (LabEx DAMAS)

References

[1] R.Z Valiev, T.G Langdon, Metall Mater Trans A 42 (2011) 2942–2951

[2] A Azushima, R Kopp, A Korhonen, D.Y Yang, F Micari, G.D Lahoti, P Groche, J Yanagimoto, N Tsuji, A Rosochowski, A Yanagida, CIRP Ann Manuf Technol 57 (2008) 716–735

[3] Y Estrin, A Vinogradov, Acta Mater 61 (2013) 782-817

[4] R.Z Valiev, R.K Islamgaliev, I.V Alexandrov, Prog Mater Sci 45 (2000) 103–189

[5] V.M Segal, Mater Sci Eng A 197 (1995) 157–164

[6] M Furukawa, Z Horita, M Nemoto, T.G Langdon, J Mater Sci 36 (2001) 2835-2843 [7] T.G Langdon, J Mater Sci 42 (2007) 3388–3397

[8] A.P Zhilyaev,T.G Langdon, Prog Mater Sci 53(2008) 893–979

[9] L.S Tóth, M Arzaghi, J.J Fundenberger, B Beausir, O Bouaziz, R Arruffat-Massion, Scripta Mater 60 (2009) 175–177

[10] N Pardis,R Ebrahimi, Mater Sci Eng A 527 (2009) 355–360

[11] N Pardis, R Ebrahimi, Mater Sci Eng A 527 (2010) 6153–6156

[12] J Richert, M Richert, Aluminium 62 (1986) 604–607

[13] N Pardis, B Talebanpour, R Ebrahimi, S Zomorodian, Mater Sci Eng A 528 (2011) 7537–7540

[14] A Zangiabadi, M Kazeminezhad, Mater Sci Eng A 528 (2011) 5066–5072

[15] A Zangiabadi, M Kazeminezhad, Comput Mater Sci 59 (2012) 174-181

[16] K Mills, J.R Davis, Metals Handbook, vol 9, sixth printing, ASM International,

Materials Park, OH, 1995

[17] Ch Xu, Z Száraz, Z Trojanová, P Lukácˇ, T.G Langdon, Mater Sci Eng A 497 (2008) 206–211

[18] W Hosford, R Caddell, METAL FORMING Mechanics and Metallurgy, third ed.,

Cambridge University Press, Cambridge, 2007

[19] L.S Toth, Y Estrin, R Lapovok, C.F Gu, Acta Materialia, 58 (2010) 1782-1794

[20] L.S Toth, B Beausir, C.F Gu, Y Estrin, N Scheerbaum, C.H.J Davies, Acta Mater 58 (2010) 6706-6716

[21] T Sakai, A Belyakov, H Miuray., Metall Mater Trans A 392 (2008) 2206-2214

[22] T Sakai, A Belyakov, R Kaibyshev, H Miura, J.J Jonas, Prog Mater Sci 60 (2014) 130–