Author's Accepted Manuscript

Tensile behavior and deformation mechan-

isms of ultrafine-grained aluminum processed
using equal-channel angular pressing
K.V. Ivanov, E.V. Naydenkin

www.elsevier.com/locate/msea

PII: S0921-5093(14)00405-5
DOI: http://dx.doi.org/10.1016/j.msea.2014.03.114
Reference: MSA30963

To appear in: Materials Science & Engineering A

Received date: 10 January 2014

Revised date: 12 March 2014
Accepted date: 25 March 2014

Cite this article as: K.V. Ivanov, E.V. Naydenkin, Tensile behavior and
deformation mechanisms of ultrafine-grained aluminum processed using
equal-channel angular pressing, Materials Science & Engineering A, http://dx.doi.
org/10.1016/j.msea.2014.03.114

This is a PDF file of an unedited manuscript that has been accepted for
publication. As a service to our customers we are providing this early version of
the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting galley proof before it is published in its final citable form.
Please note that during the production process errors may be discovered which
could affect the content, and all legal disclaimers that apply to the journal
pertain.
Tensile behavior and deformation mechanisms of ultrafine-grained aluminum processed

using equal-channel angular pressing

K.V. Ivanov*, E.V. Naydenkin

Institute of Strength Physics and Materials Science of Siberian Branch of Russian

Academy of Sciences, pr. Academichesky 2/4, 634021, Tomsk, Russia

*Corresponding author. E-mail: ikv@ispms.tsc.ru. Tel. +73822286961

Abstract

The structure and deformation behavior of ultrafine-grained aluminum processed by

equal channel angular pressing through plastic flow at room temperature in a strain rate

interval of 1.0×10-5 to 8.6×10-3 s-1 have been studied. The contribution of grain

boundary sliding to the overall deformation () was calculated using displacement of

grains relative to each other in local areas of a gage length. The strength characteristics

and tendency toward necking diminish when decreasing the strain rate, while elongation

up to failure and  increase. Improving the ductility at low strain rates should be related

to the increase in strain rate sensitivity caused by enhancing , reaching 45% in the

local areas of the neck and totaling 72% in the local areas of uniformly elongated

portion. The relatively low value of the strain rate sensitivity (m=0.08) is due to the

heterogeneity of grain structure in UFG aluminum.

Keywords: electron microscopy; mechanical characterization; aluminum;

ultrafine-grained structure; grain boundary sliding

 1. Introduction

Refining grain structures is a promising way to enhance mechanical properties of

materials. The techniques used to produce ultrafine-grained (UFG, d<1000 nm) and

nanocrystalline (d<100 nm) grain sizes have been developed until recently. The

methods involving severe plastic deformation are of particular interest because they

enable the production of bulk billets free of residual porosity and contamination [1].

Refining the grain size through severe plastic deformation increases the ultimate and

yield stresses. However, ductility drops significantly, restricting the applications of

these UFG materials. To improve their mechanical properties, understanding and

controlling the deformation mechanisms must be achieved by manipulating structural

parameters or variations in the strain conditions.

Intragrain dislocation slip and twinning are the dominant mechanisms of plastic

deformation for coarse grained metallic materials at moderate temperatures ( < 0.5m,

where Tm is the melting temperature). The role of diffusion-controlled mechanisms,

such as grain boundary sliding (GBS) and diffusion creep, is negligible [2].

Concurrently, the direct and indirect data on the essential increase in grain boundary

diffusivity in UFG metals are available in the literature. For example, the grain

boundary diffusion coefficient in UFG nickel is several times higher than for its coarse

grained counterparts [3,4]. The temperature interval of activation of some diffusion-

controlled processes in UFG materials shifts to lower temperature due to the increase of

grain boundary diffusivity [1,5,6].

The role of dislocation-related deformation mechanisms should be declined in

UFG structures due to the difficulties during the generation and movement of the

dislocation and to the increase of grain boundary diffusivity. In fact, computer

2
simulation and experimental studies show that the dislocation slip changes to GBS [7],

grain rotation [8], diffusion creep [9] and grain boundary migration (grain growth) [10]

in nanocrystalline materials. The grain size in UFG materials processed using severe

plastic deformation exceeds that in nanocrystalline materials. Nevertheless, GBS is

experimentally observed in UFG FCC metals under tension, compression and

indentation at room temperature [11-14], corresponding to 0.31Tm for aluminum,

0.22Tm for copper and 0.17 Tm for nickel. However, the contribution of GBS to the

overall deformation in these works was estimated under fixed strain conditions.

Therefore, the effect of the test conditions on the role of the different deformation

mechanisms has not been established. Consequently, the aims of the present work are as

follows: (i) investigate the peculiarities of the UFG structure using aluminum as an

example; (ii) study its deformation behavior at room temperature depending on the

strain rate; (iii) establish the effect of the strain rate on the contribution of the GBS to

the overall deformation and determine the role of this deformation mechanism under

various strain conditions.

2. Experimental

2.1. Structure investigation and mechanical testing

The material used in this study was pure aluminum (99.99%). The UFG structure

was formed by equal-channel angular pressing (ECAP) at room temperature. The ECAP

die channel had a square cross section with 10 mm sides. The internal angle ())

between the die channels was 90°, and the angle () subtended by the outer arc of

curvature at the intersection between the two channel sections was 37°. The billet was

pressed 8 times following the Bc route [15].

3
 Flat dog-bone tensile samples with a gauge size of 8×2.5×1 mm3 were electro-

discharge machined in the longitudinal horizontal section of the UFG aluminum billet.

Because structural parameters may vary in the different parts of the billet [16,17], the

structure of each test sample was studied using electron back-scatter diffraction analysis

(EBSD) to identify and exclude samples with a badly refined structure. One of the flat

surfaces of the test samples was ground to remove approximately a 50 Pm thickness if

the surface layer and mechanically polished with a succession of abrasives concluding

with a 1 Pm diamond suspension. The final electropolishing was accomplished at 30 V

for 10-20 s using a 10% HClO4 + 90% CH3OH solution cooled to 243 K.

To measure the local strain of a region in the sample, the surface was marked with

a line of circles 50 Pm in diameter and ~1 Pm deep using a focused ion beam. The

distance between the circles was 250 Pm. The tensile test resulted in the distortion of

the circles and increasing the distance between them, as easily measured by scanning

electron microscopy (SEM), giving the nearby strain value. Knowledge of the local

strain is important to evaluate the contribution of GBS to the overall deformation

correctly because it differs significantly from the total elongation of the test sample due

to the very pronounced localization of plastic deformation. The structure of the prepared

samples was investigated using transmission electron microscopy (TEM), SEM and

EBSD analysis.

The TEM foils were prepared by jet electropolishing in an electrolyte containing

6% HClO4 + 94% CH3OH at 243 K and 30 V. A Jeol 2100 microscope operated at 200

kV was used for conventional bright- and dark-field imaging. The size of the elements

of the grain-subgrain structure was measured as a maximal distance between the nearest

visible boundaries.

4
 A Quanta 200 3D scanning electron microscope with a tungsten hot cathode

operated at 30 kV was used for SEM studies and as an orientation probe during the

EBSD investigations. The dimension of the EBSD scanned areas was 50×50 Pm. A

hexagonal scanning grid with a step size of 0.3 Pm was applied. The grain/subgrain was

a region containing more than one scan point inside which the disorientation angle

between the neighboring points did not exceed the so-called tolerance angle. The

tolerance angle was fixed at 15° and 2° for grains and subgrains, respectively. The

grain/subgrain size was defined as the diameter of a circle having the same area as the

area of the grain/subgrain. All of the structural parameters obtained by EBSD were

averaged using the following formula:

¦S d i i
d ! i 1
N
(1)
¦S
i 1
i

where Si and di are the area and the diameter of the i-th grain, respectively. This

formula accounts for the area covered by grains of a definite size, making it more

significant for UFG materials processed by ECAP with a wide grain size distribution. It

should be noted that the calculation of (1) generally gives a higher average value than

that calculated as the simple average. Boundaries with misorientation less than 2° were

not considered. Low angle boundaries were regarded to have misorientation from 2 –

15°. Boundaries with a misorientation angle greater 15° were considered high angle

grain boundaries.

The tensile tests were carried out using an Instron 3369 testing machine at room

temperature. The initial strain rates were 8.6×10-3, 1.0×10-4, 3.1×10-5 and 1.0×10-5 s-1.

5
The yield stress y was the stress corresponding to the plastic strain (=0.2%) and the

ultimate stress (u) met the maximum on the stress – strain curves.

The strain rate sensitivity m was calculated using the following:

m=ln/ln H (2)

where  is the ultimate or yield stress, and H is the strain rate.

2.2. Calculation of GBS contribution to the overall deformation

The GBS contribution to the overall deformation () was measured in local areas

containing visible signs of GBS. Because there were areas exhibiting no GBS-related

deformation relief, the GBS contribution of to the overall deformation of the whole

sample was not estimated. The approach proposed by Langdon [18-20] was used. The

occurrence of GBS will lead to the occurrence of offsets in three mutually perpendicular

directions at any selected boundary. Fig. 1 a shows the scheme of the sample’s surface

cross-section oriented along the tensile direction. Two orthogonal displacements lie in

the plane of the cross-section (u and v) due to the occurrence of GBS between grains 1

and 2 under the action of an applied stress (). The third displacement (w) is

perpendicular to the section and is not shown. The strain due to GBS can be estimated

from individual measurements of one or more of these three separate displacements

[20]. The transversal component of the sliding vector perpendicular to the sample

surface v was the most convenient. v was visualized and measured in SEM on cross-

sections of the surface prepared using FIB (see the insert in Fig. 1 a). First, a platinum

stripe 40×1.5×1.5 m in size was deposited on the area of interest by means of the FIB

assistant gas injection system to improve the SEM contrast and to prevent surface

damage during the mill. Next, the cross-section of the interface platinum stripe - surface

6
was constructed. Because one took a side view, a profile of the surface was clearly

visible and measurable (Fig. 1 b). Three stripes in the each studied local area were

cross-sectioned. The number of v measurements depended on the relief: generally 8 –

10 on each cross-section. The sliding strain was calculated: [18]

Hgbs=kn<v>, (3)

where k 1.5 is a constant, n is the number of boundaries with displacement per

unit length and <v> is the average value of v. The contribution of GBS to the overall

deformation was expressed as =Hgbs/H, where H is a strain in the studied local area.

The relative error of the  measurement was evaluated using the formula

=((n/<n>)2+(v/<v>)2+(/)2))1/2,

where n is the absolute error of the n measurement, <n> is the simple average of

n, v is the absolute error of the v measurement and / is the relative error of the

strain measured in a local area. / was 0.1; n and v were calculated according to

standard statistical methods.

3. Results and discussion

3.1. Microstructure

The TEM studies reveal the essential refinement of the microstructure of

aluminum due to ECAP (Fig. 2). The elements of the grain-subgrain structure are

equiaxed, varying in size from 0.5 to 4 Pm. The extinction contours are observed inside

of them, suggesting the presence of internal stresses. Most of the dislocations are

organized in the form of walls, pile-ups and networks, while others are individual. The

dislocation density inside the grain-subgrain structure elements is rather low. Fig. 2

shows that the grain/subgrain boundaries and triple junctions are sources of internal

7
stresses. The structural features obtained by routine TEM are the same in the specimens

obtained from different parts of the ECAP processed billet. No any structural

heterogeneity is found. The above results agree with the experimental data on the

structure of metals processed by ECAP that are widely available in literature, e.g.,

[12,21-24]. TEM is a very local method, and it is difficult to study large areas.

Moreover, separating the areas containing mainly low angle boundaries from those

containing mainly high angle ones requires craftsmanship because they look similar

during TEM studies. It is hard to measure the dimensions of the areas. These limitations

must be noted for TEM.

A careful EBSD investigation reveals that the subgrain size varies from 0.5 –

7 m, averaging 2.3 – 3.2 m (Fig. 3 a, b, e, f). The subgrains below 4 m cover most

of the scan area. The subgrain size distributions are the same and the average subgrain

size varies insignificantly in the EBSD scans obtained in the different regions of the test

samples (compare the subgrain size distributions in Fig. 3 b, f). The relatively

homogeneous subgrain structure is typical for UFG aluminum processed by ECAP

[17,25,26].

The grain structure is more complicated. The grain size distributions obtained in

the different regions of test sample show the different fractions of the coarse grains (see

Fig. 3 c, g). The coarse grains are grains larger than the largest subgrain: 7 m. In the

scans with a well-refined grain structure the area fraction for the coarse grains is 10 –

15%, the average grain size is typically 4 – 5 m, the maximum grain size is 10 m

and the fraction of high angle grain boundaries reaches 70% (Fig. 3 d). Some scans

displayed a poorly refined grain structure where the area fraction of the coarse grains, as

well as the average and the maximum grain size, rise; meanwhile, the fraction of the

8
high angle grain boundaries decreases. For example, the worst refined grain structure in

the test samples has 40% of coarse grains with average and maximum grain sizes being

8.8 and 20 m, respectively. The fraction of high angle grain boundaries decreased

down to 55% in these areas (Fig. 3 g, h). If the grain structure parameters in a sample

were worse, the sample was excluded from the mechanical testing.

The coarse grains usually form elongated bands tens of micrometers in width and

more than the size of EBSD scans (50 m) in length. The presence of such bands is

indirectly confirmed by literature data based on EBSD grain maps [26,27]. The bands in

UFG aluminum with identical texture components have been described in [28,29].

These texture bands are similar in size compared to the bands of coarse grains, and they

most likely have a common origin. Therefore, the grain structure of UFG aluminum is

heterogeneous. The average grain size and fraction of the high angle grain boundaries

depends on the fraction of coarse grains in the definite EBSD scan. This heterogeneity

can affect the deformation behavior of the UFG material.

3.2. Deformation behavior

The engineering stress – strain curves obtained at different strain rates are shown

in Fig. 4. Apparently, the high values for the ultimate and yield stresses, the relatively

low elongation up to failure, the modest value of the strain hardening, approach of the

ultimate stress and the onset of the deformation localization at a low plastic strain are

characteristic features of the deformation behavior of UFG aluminum under tension.

These values are typical for UFG materials and related to the fact that dislocation

density saturates due to the dynamic recovery or the annihilation of dislocations into the

grain boundaries [30]. These features are most pronounced when testing the maximal

9
strain rate. As the strain rate decreases, the difference between the ultimate and yield

stresses increases, and the strain up to the onset of necking and the total elongation

rises.

The distribution of plastic strain in the gage length of the test samples reveals that

at 8.6×10-3 s-1, practically the all strain is localized in a neck, while 3/4 of the gage

length remains strainless after the tension test. The decreased strain rate increases the

uniform elongation and reduces strain in the neck. A maximal uniform elongation of

12 – 25% is observed at 1.0×10-5 s-1. The, the decreased strain rate of UFG aluminum

decreases the tendency toward necking (Fig. 5).

Dependence of the ultimate and yield stresses on the strain rate in logarithmic

coordinates are plotted in Fig. 6. As observed, the dependence is not linear. The strain

rate sensitivity (m) varies from 0.02 – 0.04 at high strain rates, rising to 0.08 when

decreasing the strain rate. The increase in m may be related to the development of GBS

and the enhancement of its contribution to the overall deformation. However, the m

value remains lower than that corresponding to GBS (mGBS=0.5).

Enhancing the uniformity of elongation and improving the ductility by decreasing

the strain rate were demonstrated previously for UFG copper [31], aluminum [32,33]

and aluminum alloys [34-36]. The effect should be related to the increased value of m

according to the Hart’s criterion [31]. However, there was no discussion in [31]

regarding the physical reasons of the m rising in UFG metals by decreasing the strain

rate. In [35], it was assumed that the ductility rises due to the interplay of GBS and the

shear banding in commercial aluminum alloy 6082. In turn, in shear bands, GBS occurs.

However, the authors of [32] suppose that the m value is rather low for the tension of

UFG aluminum at room temperature, making GBS an implausible explanation for the

10
enhanced ductility at low strain rates. When accounting for the above, the measurement

of the contributions of GBS to the overall deformation will reveal the deformation

mechanisms that act during the tension of UFG metals at different strain rates.

3.3. Deformation relief

The SEM investigation shows that the deformation relief arising on the polished

surfaces of the tested samples during deformation depends on the local strain in the site

of observation at all of the studied strain rates. As mentioned above, 3/4 of the gage

length of the sample remains strainless after the test at 8.6×10-3 s-1; and the other 1/4

represents the neck. No deformation relief is apparent on the strainless part of the gage.

In the local region strained to 10–20%, bands of localized deformation (BLD) appear on

the surface. They are visible as relatively coarse surface folds perpendicular to the

tensile direction that are 2–4 m wide and several hundred microns in length. The

distance between BLS is 10–20 m or more (Fig. 7 a). The relief characteristic to GBS

is observed inside BLD and the adjacent areas (see the area rounded by the white oval in

Fig. 7 b). No GBS relief is apparent between BLS in this region. The difference

between the BLD and the micro shear bands described in [35] should be noted. First, the

scratches bend but do not break in BLD, suggesting BLD are not related to GBS.

Second, they are different in size. Third, the number of the micro shear bands on a

square unit rise when decreasing a stain rate and vice versa for BLD.

The long BLD are divided into parts tens of microns in length in the regions

locally strained more than 30%. The number and width of the BLD increase. They cover

practically all of this part of the gage. The network of grain boundaries due to GBS is

visible between BLD; the more strain, the more GBS will occur. Signs of intragrain

11
dislocation slip, arising as long straight lines and progressing from boundary to

boundary, can be observed inside grains coarser than 5 m (Fig. 7 c).

Deformation relief arising on the sample tested at H =10-4 s-1 differs significantly

from the relief described above. The network of boundaries induced by GBS is clearly

observed in the uniform elongated gage length of the sample (outside neck,  8 – 10%,

Fig. 8 a). In the neck region the deformation relief changes: areas containing BLD and

containing no traces of GBS appear (Fig. 8 b). These areas have a size of 100 m in

width and several hundred microns in length. The size and the shape of the areas are

similar to those of the coarse grained bands observed in EBSD maps and containing

mainly low angle boundaries. Because GBS generally occurs on high angle grain

boundaries, BLD should develop mainly in the areas with a poorly refined grain

structure.

Further decreases in the strain rate down to H =10-5 s-1 do not qualitatively change

the deformation relief. However, the GBS-related relief is more pronounced in the

uniform elongated part and BLD-containing areas decrease in size. In the necking

region (>40%) there are no areas without GBS-related relief.

3.4. GBS and its contribution to the overall deformation

The dependence of the GBS contribution to the overall deformation () versus the

strain in a local area is presented in Fig. 9.  is the upper limit of the contribution

because there are areas where BLD develops and GBS does not. For deformation at

H =8.3×10-3 s-1, the most of the gage length remains strainless: =0. The contribution is
measurable in the neck area only: =19% in the area locally strained to 34%. At

H =1.0×10-4 and 3.1×10-5 s-1,  20% in the area of necking. In the uniform elongated
parts  reaches 27 and 72%, respectively. At a strain rate of H =1.0×10-5 s-1 the maximal

12
value of  is close to the head of the test sample (=41%, =10%), while in the necking

area, it is close to the break surface (=45%, =44%). In the uniform elongated part, 

decreases while remaining rather high,  =31%. Therefore,  varies from 20 to 70% in

the interval of the studied strain rates. Note the unexpected greater  obtained in the

area locally strained to 8% at a higher strain rate of H =3.1×10-5 s-1 relative to that at

H =1.0×10-5 s-1. This is probably due to the grain structure heterogeneity of UFG
aluminum, namely, to the measurement of  in unfavorable local area with worse-

refined grain structure in the latter case. It was shown earlier that the contribution of

GBS to the overall deformation under the tension of the UFG aluminum at room

temperature and H =8.3×10-3 s-1 reached 24% [14]. For UFG aluminum,  reached 40 –

70% under indentation [13]. =25% for the compression of UFG copper at room

temperature. Evidently, the contribution of GBS to the overall deformation agrees with

the experimental data of the other researchers.

GBS may result in the structural changes of a material, therefore the structure of

the sample tested at H =1.0 10-5 s-1 (it showed the maximum uniform elongation and the

maximum  in the most local areas) was studied by EBSD. Fig. 10 demonstrates inverse

pole figure with superimposed boundary map, the respective grain size distribution and

the boundary distribution on misorientation angle obtained in the local area strained to

19% and exhibiting strongly pronounced GBS-related deformation relief. The following

features should be noted: first, there are no coarse grains in this area. The average grain

size is equal to 3.9 m and the fraction of high angle grain boundaries amounts to 72%.

Thus, GBS really occurs predominantly in the areas with well-refined grained structure.

Second, evidently from Fig. 10 that grains don’t grow during GBS. Grain growth is

frequently observed during strain of UFG materials, e.g. [12]. But the structure

13
evolution was studied by TEM, consequently, the growth of the grain-subgrain structure

elements rather than grains was found in [12]. The evolution of grain structure during

deformation of UFG metals is not established well now and requires extra study. Third,

the shape of the grains is close to the equiaxed one suggesting the considerable

contribution of GBS to the overall deformation.

As mentioned above, contradictory opinions exist in the literature regarding the

role of GBS when enhancing the strain rate sensitivity and improving the ductility of the

UFG metals by decreasing the strain rate. The present data suggest that the contribution

of GBS to the overall deformation increases when decreasing the strain rate, reaching a

significant value in local areas (up to 40 – 70% including in the uniform elongated

regions). Therefore, GBS is very important. However, hypothetically bands with

relatively coarse grains and low fraction of high angle grain boundaries exist where

GBS occurs only in the neck region and at lowest strain rate. The GBS contribution to

the overall deformation of the whole sample decreases, possibly explaining the

relatively low value of m. To improve the ductility of the UFG materials by increasing

the strain rate sensitivity and the contribution of GBS to the overall deformation, a

homogeneous grain structure is required. This conclusion is indirectly confirmed by the

studies of GBS and m in Al-30Zn alloy [34]. The contribution of GBS to the overall

deformation in the alloy was 40 – 60% and m 0.25 at 10-4 s-1. The essentially increased

value of m in the alloy relative to the pure aluminum with close values of  may be

attributed to the fact that in alloys, the grain structure is more homogeneous [1].

However, this suggestion requires further experimental validation.

14
 4. Conclusions

The structure, deformation behavior and contributions of grain boundary sliding to

the overall deformation occurring by the plastic flow of ultrafine-grained aluminum

processed by ECAP at room temperature in a wide strain rate interval have been

studied.

1. EBSD indicates that the grain structure of UFG aluminum is heterogeneous:

there are areas with a well-refined grained structure, a relatively small grain size, a

narrow grain size distribution and a high fraction of high angle grain boundaries (up to

70%). In the other areas (tens of microns in size), there is a significant area fraction of

coarse grains with a decreased fraction of high angle grain boundaries. The above well-

and worse grain refined areas look the same from the standpoint of EBSD sugrain maps

and ordinary TEM studies.

2. The mechanical behavior of UFG aluminum exhibits several characteristic

features compared to the coarse grained counterpart: high values of ultimate and yield

stresses, relatively low elongation up to failure, a modest value of strain hardening,

approach of the ultimate stress and the onset of deformation localization with little

plastic strain. The strength characteristics decrease, the strain hardening changes

insignificantly and elongation up to failure rises by decreasing the strain rate. The latter

effect should be explained by the increased strain rate sensitivity; this sensitivity is due

to the increased contribution of GBS to the overall deformation at a low strain rate.

3. The deformation relief study reveals that intragrain dislocation slip in relatively

coarse grains and the grain boundary sliding of ultrafine grains are involved in the

plastic flow of UFG aluminum. Direct measurements of the contribution of grain

15
boundary sliding to the overall deformation in local areas show that it increases by

decreasing the strain rate. The upper limit of the contribution reaches 45% in the

necking area and 72% in the area of the uniform elongation.

4. The value for the strain rate sensitivity remains low, even at low strain rates,

despite the significant contribution of grain boundary sliding to the overall deformation.

This effect is most likely related to the presence of the worse grain refined areas where

grain boundary sliding is limited.

Acknowledgments

The work was carried out within the framework of project No. III.23.2 of the

Siberian Branch of the Russian Academy of Sciences. The financial support of the

Ministry of Education and Science of the Russian Federation (Contract No.

02.G25.31.0063) is acknowledged.

References

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

189.

[2] H.J. Frost, M.F. Ashby, Deformation-Mechanism Maps, Pergamon Press,

Oxford,1982.

[3] S.V. Divinski, G. Reglitz, H. Rösner, Y. Estrin, G. Wilde, Acta Mater. 59 (2011)

1974–1985.

[4] Yu.R. Kolobov, G.P. Grabovetskaya, M.B. Ivanov, A.P. Zhilyaev, R.Z. Valiev,

Scripta Mater. 44 (2001) 873–878.

[5] Yu.R. Kolobov, G.P. Grabovetskaya, K.V. Ivanov, M.B. Ivanov, Interface Sci. 10

(2002) 31–36.

16
[6] B.B. Straumal, B. Baretzky, A.A. Mazilkin, F. Phillipp, O.A. Kogtenkova, M.N.

Volkov, R.Z. Valiev, Acta Mater. 52 (2004) 4469–4478.

[7] H. van Swygenhoven, P. Derlet, Phys. Rev. B 64 (2001) 224105.

[8] N.I. Noskova, J. Alloys Compounds, 434–435 (2007) 307–310.

[9] V. Yamakov, D. Wolf, S.R. Phillpot, H. Gleiter, Acta Mater. 50 (2002) 61ದ73.

[10] A.J. Haslam, D. Moldovan, V. Yamakov, D. Wolf, S.R. Phillpot, H. Gleiter, Acta

Mater. 51 (2003) 2097–2112.

[11] A. Vinogradov, S. Hashimoto, V. Patlan, K. Kitagawa, Mater. Sci. Eng. A 319-321

(2001) 862ದ866.

[12] R.Z. Valiev, E.V. Kozlov, Yu.F. Ivanov, J. Lian, A.A. Nazarov, B. Baudelet, Acta

Metall. Mater. 42 (1994) 2467ದ2475.

[13] N.Q. Chinh, P. Szommer, T. Csanadi, T.G. Langdon, Mater. Sci. Eng. A 434

(2006) 326–334.

[14] K.V. Ivanov, E.V. Naidenkin, Scripta Mater. 66 (2012) 511–514.

[15] R.Z. Valiev, T.G. Langdon, Prog. Mater. Sci. 51 (2006) 881–981.

[16] K.V. Ivanov, E.V. Naidenkin, Russian Phys. J. 52 (2009) 1030–1035.

[17] K.V. Ivanov, E.V. Naidenkin, Rev. Adv. Mater. Sci. 25 (2010) 176–182.

[18] T.G. Langdon, Metall. Trans. 3 (1972) 797–801.

[19] R.L. Bell, C. Graeme-Barber, T.G. Langdon, Trans. AIME 239 (1967) 1821–1824.

[20] T. G. Langdon, J. Mater. Sci. 41 (2006) 597–609.

[21] Y. Iwahashi, Z. Horita, M. Nemoto, T.G. Langdon, Acta Mater. 46 (1998) 3317–

3331.

[22] C.P. Chang, P.L. Sun, P.W. Kao, Acta Mater. 48 (2000) 3377–3385.

17
[23] O.V. Mishin, D. Juul Jensen, N. Hansen, Mater. Sci. Eng. A 342 (2003) 320–328.

[24] M. Cabibbo, W. Blum, E. Evangelista, M.E. Kassner, M.A. Meyers, Metall. Mater.

Trans. A 39A (2008) 181–189.

[25] S.D. Terhune, D.L. Swisher, K. Oh-ishi, Z. Horita, T.G. Langdon, T.R. Mcnelley,

Metall. Mater. Trans. A 33A (2002) 2173–2184.

[26] I. Saxl, V. Sklenika, L. Ilucová, M. Svoboda, J. Dvorák, P. Král, Mater. Sci. Eng.

A 503 (2009) 82–85.

[27] I. Saxl, A. Kalousová, L. Ilucová, V. Sklenika, Mater. Charact. 60 (2009) 1163–

1167.

[28] A.P. Zhilyaev, D.L. Swisher, K. Oh-ishi, T.G. Langdon, T.R. McNelley, Mater.

Sci. Eng. A 429 (2006) 137–148.

[29] K. Oh-ishi, A.P. Zhilyaev, T.R. McNelley, Mater. Sci. Eng. A 410-411 (2005)

183–187.

[30] M.A. Meyers, A. Mishra, D.J. Benson, Prog. Mater. Sci. 51 (2006) 427ದ556.

[31] Y.M. Wang, E. Ma, Acta Materialia 52 (2004) 1699–1709.

[32] H.W. Höppel, J. May, M. Göken, Adv. Eng. Mater. 6 (2004) 781ದ784.

[33] J. May, H.W. Höppel, M. Göken, Mater. Sci. Forum. 503-504 (2006) 781–786.

[34] R.Z. Valiev, M.Yu. Murashkin, A. Kilmametov, B. Straumal, N.Q. Chinh, T.G.

Langdon, J. Mater. Sci. 45 (2010) 4718–4724.

[35] I. Sabirov, Yu. Estrin, M.R. Barnett, I. Timokhina, P.D. Hodson, Acta Mater. 56

(2008) 2223–2230.

[36] I. Sabirov, M.R. Barnett, Yu. Estrin, P.D. Hodgson, Scripta Mater. 61 (2009) 181–

184.

18
 Captions for figures

Fig. 1. (a) A two-dimensional scheme of GBS between grains A and B adapted from

[20]. u and v denote the offsets of the sliding vector S. (Insert) v was measured using

SEM. The tilt of the sample in the SEM is accounted for. (b) The cross-section of the

interface “platinum stripe – surface” of the tested sample. The white arrows indicate

steps on grain boundaries due to GBS.

Fig. 2. Bright field TEM image of the UFG aluminum structure. Extinction contours are

concentrated around the triple junction in the center of the image, suggesting the

presence of internal stresses.

Fig. 3. Typical EBSD data obtained for a test sample. (a, e) The inverse pole figures

with a superimposed boundary map: black and white lines correspond to high- and low

angle boundaries, respectively. (b, f) Subgrain size distributions. (c, g) Grain size

distributions. (d, h) Distributions of boundaries on misorientation angle. (a – d)

correspond to the area with well-refined grain structure, and (e – h) correspond to the

area with the worst-refined grain structure. Note the similar subgrain size distributions

in (b, f) and the essentially different grain size distributions in (c, f).

Fig. 4. Engineering stress – strain curves obtained for UFG aluminum at different strain

rates.

Fig. 5. Distribution of strain in the local areas on the gage length calculated as a

distortion of the mark circles. The maximum value in the neck is estimated using the

distance between the nearest visible mark circles.

Fig. 6. The ultimate and yield stress versus strain rate obtained from the tensile tests.

Note the greater slope of the dependence at lower strain rates.

19
Fig. 7. Deformation relief on the surface of the sample tensioned at 8.6×10-3 s-1. (a, b)

=12%. (c) = 34%. The white thick arrows mark the bands from localized deformation.

GBS in the vicinity of the band is surrounded by the white oval in (b). Straight lines

from boundary to boundary due to intragrain dislocation slip are visible inside the

coarse grains in (c).

Fig. 8. Deformation relief on the surface of the sample tensioned at 1.0×10-4 s-1. (a)

=8%. (b) =26%. The area where preferably BLD appears is surrounded by the white

oval in (b). The extensive area where GBS occurs is marked, respectively.

Fig. 9. The contribution of GBS to the overall deformation in the local area versus the

strain in this area obtained for the samples tested at different strain rates.

Fig. 10. EBSD data obtained in the uniform elongated local area of the tested sample

( H =1.0 10-5 s-1, =19%) exhibiting GBS. (a) The inverse pole figure with superimposed

boundary map. Marking of boundaries is the same as in Fig. 3. (b, c) Grain size

distribution and distribution of boundaries on misorientation angle.

20
Figure(s) 1
Figure(s) 2
Figure(s) 3a and 3e
Figure(s) 3a and 3e black and white
Figure(s) 3b, 3c, 3d, 3f, 3g and 3h
Figure(s) 4
Figure(s) 4 black and white
Figure(s) 5
Figure(s) 5 black and white
Figure(s) 6
Figure(s) 7a, 7b and 7c
Figure(s) 8a and 8b
Figure(s) 9
Figure(s) 9 black and white
Figure(s) 10a
Figure(s) 10a black and white
Figure(s) 10 b c