Materials Science & Engineering A 563 (2013) 193–198
Contents lists available at SciVerse ScienceDirect
Materials Science & Engineering A
journal homepage: www.elsevier.com/locate/msea
TEM analysis and determination of dislocation densities in nanostructured
copper tube produced via parallel tubular channel angular pressing process
G. Faraji a,b,n, M.M. Mashhadi a, A.R. Bushroa b, A. Babaei a
a
b
Department of Mechanical Engineering, University of Tehran, Tehran 11155-4563, Iran
Centre of Advanced Manufacturing and Material Processing, Level 8, Engineering Tower, Faculty of Engineering, University of Malaya, 50603 Kuala Lumpur, Malaysia
a r t i c l e i n f o
abstract
Article history:
Received 31 October 2012
Received in revised form
17 November 2012
Accepted 19 November 2012
Available online 27 November 2012
Parallel tubular channel angular pressing (PTCAP) is a recently developed novel intense plastic
deformation method appropriate for fabrication ultrafine-grained and nanostructured cylindrical tubes.
In the present work a commercially pure copper was processed via multi-pass PTCAP and the effects of
number of passes on grain refinement and the dislocation density were studied. TEM analysis showed
that after first pass elongated subgrains with interior tangled dislocations were formed. After pass
number two the density of interior dislocations through the elongated grains was decreased. In the next
stages of deformation, at pass number three, elongated grains almost disappeared and equiaxed grains
with grain size about 150 nm are formed as a result of dynamic recovery. The dislocation densities were
measured by hardness indentation size effect using the Nix–Gao model. The results showed that
increase in the number of PTCAP passes leads to decrease in the dislocation densities. The dislocation
density is decreased to 2.48 109 cm 2 after fourth passes from 18.1 109 cm 2 after first pass. TEM
results verified calculated values from the Nix–Gao model. Microhardness of the PTCAP processed tube
through four passes increased to 142 HV from initial value of about 62 HV. Also, significant increase
takes place after single pass and in the next passes the hardness value is saturated.
& 2012 Elsevier B.V. All rights reserved.
Keywords:
Parallel tubular channel angular pressing
Nanostructured tube
TEM
Dislocation density
Hardness
1. Introduction
It is well-founded that there are superior mechanical and
physical properties in ultrafine grain (UFG) and nanostructured
(NS) materials [1]. Improving of the material properties using
grain refinements via severe plastic deformation (SPD) has been
interested in past two decades [2]. In recent decades different SPD
methods were developed for producing UFG and NS bulks [3–7].
The ultra-high plastic strain imposed during SPD subdivides
coarse grains into UFGs, and simultaneously creates a high
density of dislocations on the order of 1014–1016 m 2 [8–10].
Tubular components have more application in different industries
but few efforts have been done for producing UFG and NS tubes
employing SPD techniques. An especially attractive method,
tubular channel angular pressing (TCAP) based on equal channel
angular pressing (ECAP), developed by Faraji et al. [11–13]. They
applied this technique to AZ91 magnesium alloy and reported
more grain refinement and significant improvement in the
mechanical properties [11]. They also investigated the effects of
process parameters such as channel geometry [12] and channel
n
Corresponding author at: Department of Mechanical Engineering, University
of Tehran, Tehran, 11155-4563, Iran. Tel./fax: þ98 2161114033.
E-mail address: [email protected] (G. Faraji).
0921-5093/$ - see front matter & 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.msea.2012.11.065
angle [13] on plastic deformation and process load using finite
element modeling. Also, very recently they proposed a new SPD
process named PTCAP method based on TCAP suitable for producing high strength tubes [14]. They showed that obtaining
excellent strain and hardness homogeneity and also needing
60% lower process loads which are two major advantages of
PTCAP compared to TCAP. As is well-known, process load and
hardness homogeneity are two important challenges in SPD of
metallic materials, hence PTCAP is an excellent SPD method for
producing UFG and NS tubular components [14].
During PTCAP process the material is severely deformed and
UFG materials are achieved. To develop UFG structure the
presence of excess dislocations are needed [15]. As reported in
previous researches, the dislocation densities act as a determinant
factor in controlling the mechanical properties of the metals
[2,5,8,16,17]. Whereas existence of high density of dislocations
plays a main role in producing UFG materials, estimation of the
dislocation density is necessary. The dislocation density can be
measured using TEM micrographs and also using the Nix–Gao
model [18]. Graca et al. reported that it is possible to estimate the
dislocation densities by the hardness indentation size using the
Nix–Gao model [18]. In the current study a commercially pure
copper tube was processed via multi-pass PTCAP and the effects
of number of passes on grain refinement, mechanical properties
and the dislocation density were studied.
194
G. Faraji et al. / Materials Science & Engineering A 563 (2013) 193–198
Fig. 1. Principles of the PTCAP process in the (a) first and (b) second half cycles and (c) parameters.
In the PTCAP process there are two half cycles shown schematically in Fig. 1. The first punch presses the tube material into
the gap between mandrel and die including two shear zones
shown in Fig. 1(a) during the first half cycle. In this stage, the
diameter of the tube increases to gain its maximum value. Then,
the tube is pressed back using second punch in the second half
cycle (Fig. 1(b)). The tubes thickness in initial and processed
conditions is same. The process including two half cycles can be
repeated to apply a distinct strain while no variation in the cross
section area of the tube is occurred. After N passes of PTCAP
process, the total equivalent plastic strain can be estimated via
the following equation [14]:
(
eTN ¼ 2N
)
2 X
2cot ji = 2 þ ci = 2 þ ci cosec ji = 2 þ ci = 2
2
R2
pffiffiffi
þ pffiffiffi ln
3 R1
3
i¼1
ð1Þ
These formulas are different from conventional ECAP because
there are some peripheral tensile and compression strains [11,12]
in PTCAP process. From Eq. (1), the equivalent plastic strain
through 1, 2, 3 and 4 passes of PTCAP are 3, 6, 9 and 12,
respectively.
2. Experimental procedures
The samples of commercially pure copper were taken as
initial tubes of 2.5 mm in thickness, 20 mm in diameter, and
50 mm in length. Prior to PTCAP processing, tubes were heated
to the temperature of 600 1C and annealed for 2 h to get a
recrystallized homogeneous microstructure shown in Fig. 2(a).
A PTCAP die including mandrel, die and punches from H13 tool
steel was prepared and hardened to 55 HRC. Before pressing
G. Faraji et al. / Materials Science & Engineering A 563 (2013) 193–198
195
Fig. 2. Optical microstructure of annealed copper (a) bright-field TEM micrographs of the cross sectional area of PTCAP processed tube through different number of passes
(b), (e) one pass, (c), (f) two passes, (d), (g) three passes.
the tube, the die parts were lubricated with MoS2 [19]. Die
parameters including the channel
angle f1 ¼ f2 ¼ 1203 , the
3
angle of curvature c1 ¼ c2 ¼ 0 and K ¼ R2 R1 was equal to
the tube thickness (2.5 mm) are shown in Fig. 1(d). PTCAP
experiments were done in punch speed of 5 mm/min at RT.
Microhardness measurements was performed in loads ranged
from 0.1 to 5 N. The microstructure in the PTCAP processed tube
was investigated by transmission electron microscopy (TEM).
Lamella sample preparation was performed using focused ion
beam (FIB). Two trenches are milled on each side of the area
of interest until it is about 1 mm thick. The upper and lower
side of the TEM sample are then milled together with a U-cut
under the sample leaving only a small uncut portion to hold
the lamella to the main sample. The lamella is finally attached
to a TEM Cu grid using platinum deposition and the needle
is milled away. Final thinning is performed on the lamella
sample. The grain sizes were measured using the linear interpret
method.
3. Results and discussion
Fig. 2(a) shows initial recrystallized microstructure of the
unprocessed sample in which a mean grain size of 59 mm could
be seen. The microstructure evolution of pure copper tube after
PTCAP is shown in Fig. 2(b)–(g). Fig. 2(b) and (e) shows the TEM
micrograph of PTCAP processed tube in perpendicular-with-axis
direction through one pass, at equivalent strain 3, in two
different magnifications. As expected, the microstructures are
more refined, and are significantly different from the initial state
(Fig. 2(a)). TEM micrograph of first pass PTCAP processed specimen contains an elongated and subgrain structure with tangled
dislocations in which distinct regular-shaped walls was observed.
This event was obviously seen in pure copper after second pass in
the ECAP process [20]. It is also found that the subgrain size in the
microstructure of the initial tube is more coarser than that in the
PTCAP processed one. After undergoing the first pass, original
coarse grains are divided into subgrains. TEM micrographs of the
196
G. Faraji et al. / Materials Science & Engineering A 563 (2013) 193–198
Hardness (Hv)
tube performed by PTCAP process for second pass which corresponds to equivalent strain of 6 are shown in Fig. 2(c) and (f).
In this figure, even though some of the local tangled dislocations
still exist, the elongated grains with 150 nm size can be
distinguished by sharp boundaries. At this stage the density of
interior dislocations is decreased but the grains is not still free of
dislocations. This indicates that dynamic recovery (annihilation or
reorganization) occurs near these boundaries [20]. The grain
refinement process implies the arrangement and rearrangement
of dislocation cells in original grains of the pure copper [21–23].
When the number of PTCAP passes increase to three, which is at
equivalent strain 9, elongated grains are almost disappeared,
and equiaxed grains with grain size about 150 nm are formed
(Fig. 2(d) and (g)). Tendency for formation of recrystallized
equiaxed grains with sharp boundaries from elongated one when
increasing in the accumulated strains to about 10 could also be
seen in ECAP of high-purity Al [20]. Ferrasse et al. reported that
the ultra-fine grains were developed by dynamic recrystallization
through grain fragment rotation during ECAP deformation at
room temperature at higher strains [24]. Small recrystallized
grains were also observed in 99.95% copper deformed equivalent
to a strain of 10 during ECAP [25].
The Vickers microhardness changes along the tube thickness
after one to three PTCAP passes are shown in Fig. 3. Hardness
increases notably after first pass of PTCAP and then increases
gradually. Good hardness distribution was observed throughout
the thickness of the PTCAP processed tube in all PTCAP passes.
The figure shows that after third pass the microhardness is
increased to 142 HV from the primary value of 62 HV. It is also
detected that the microhardness is saturated at equivalent plastic
strain of about 6. This saturation could also be seen in ECAP
process and it was mentioned to be related with the lower bound
of the grain size which is in the range of about 200 nm [21]. It is
well known from hall–petch relationship that the hardness is
related to the grain size. Microstructural investigations establish
that the grain size is saturated at the value of 150 nm after one
pass PTCAP. The changes in hardness, from the first to third pass
of PTCAP, are in good agreement with the microstructural evolutions discussed in the previous sections. This is also verified by
the works done by other researchers shown in the chart [26–28].
The dislocation densities of the samples could be estimated
from indentation hardness measurement. Graca et al. [16,18]
mentioned that the hardness values increase as the indentation
depth decreases. The relation between the hardness value and
150
140
130
120
110
100
90
80
70
60
50
Current study
indentation depth is indicated as following equation [16]:
2
H
n 1
¼ 1 þh
H0
h
ð2Þ
where H0 is the hardness in the limitation of infinite depth (bulk
hardness), hn is a characteristic length and H is the hardness value
corresponding to indentation depth h. Fig. 4 shows the amount of
hn in the PTCAP processed specimens through 1, 2, 3 and 4 passes
which are measured to be 788.53 nm, 2958.4 nm, 4268.3 nm and
5766.4 nm respectively determined by fitting Eq. (2) to experimental data.
The correlation between the dislocation density (rs) statically
stored in the lattice and the characteristic length (hn) is as
following [8,17]:
rs ¼
3 1 tan2 y
2 f 3 bhn
ð3Þ
where b is the Burgers vector of the dislocations, rs is the density
of dislocations statistically stored in the lattice, y is the angle
between the surfaces of the material and the indenter, and f is a
correction factor for the size of the plastic zone. In this research
y ¼221, f ¼1.9 [8], and b¼0.25 nm [8,17].
The density of dislocations statistically stored in the lattice of
processed samples versus PTCAP passes calculated from Eq. (3) is
shown in Fig. 5. It is clear that an increase in the number of PTCAP
passes and accumulated strain leads to decrease in the density of
dislocations. As mentioned above, decrease in the density of
dislocations with increase in the number of passes indicates that
the dislocation annihilation by dynamic recovery occurs during
additional straining [20]. The density of dislocations in PTCAP
processed sample after one pass is about 18.1 109 cm 2 which
is 3.7 times higher than that in processed sample after two passes.
This can be verified by TEM micrographs of processed samples
through one and two passes shown in Fig. 2(b) and (c), respectively. Increase in the number of PTCAP passes from two to three
and four passes almost has not more effect on the density of
dislocations. This can also verify the hardness saturation which
was seen in Fig. 3.
Miyajima et al. estimated the density of dislocations values to
be around 1010 cm 2 in commercial pure aluminum samples with
equivalent strain up to 10 processed by accumulative roll bonding
process using the scanning transmission electron microscopy
(STEM) [29]. The present values of dislocation densities estimated
using hardness indentation have almost a good agreement with
the previously reported values by other techniques such as STEM
[29] and X-ray [30]. Barmouz et al. [8] estimated the dislocation
densities of friction stir processed (FSPed) copper sample using
the similar method used in this work to be in the range of
109 cm 2 which is lower than that of the values calculated in this
research in the lower number of passes. This may be attributed to
the heat input in the FSP process as a result of intense friction
between tool and copper sample compared to PTCAP process
done at room temperature [31].
Delta Torre et al. [ECAP]
4. Conclusions
Kadri et al
Habibi et al. (ECAR)
Kommel et al. (ECAP)
0
3
6
9
12
Equivalent plastic strain
15
18
Fig. 3. Microhardness (measured at load: 100 g and time:15 s) versus. equivalent
plastic strain in multi-passes PTCAP processed tubes compared with other
researcher’s results on SPD of copper.
A commercially pure copper was processed via multi-pass
PTCAP and the effects of number of passes on grain refinement
and accumulation of the dislocations were studied. TEM analysis
showed that after first pass elongated subgrains with interior
tangled dislocations were formed. After second pass, the density
of interior dislocations through the elongated grains was
decreased. In the next stages of deformation (third pass), elongated grains almost disappeared and equiaxed grains with grain
size about 150 nm were formed as a result of dynamic recovery.
G. Faraji et al. / Materials Science & Engineering A 563 (2013) 193–198
20
20
Pass#1
Pass#2
15
(H/H0)2
(H/H0)2
15
10
y= 788.53x+ 1
5
0
0.001
0.002
1/h(nm-1)
10
y= 2958.4x + 1
5
0
0
0.003
0
0.001
0.002
1/h (nm-1)
0.003
20
20
Pass#4
Pass#3
15
15
(H/H0)2
(H/H0)2
197
y= 4268.3x+ 1
10
5
y= 5766.4x+ 1
10
5
0
0
0
0.001
0.002
1/h (nm-1)
0.003
0
0.001
0.002
1/h(nm-1)
0.003
Dislocation density (x109cm-2)
Fig. 4. Indentation effect pictures at different loads and fitting of Eq. (2) to the microhardness data of PTCAP processed specimens through different number of passes.
25
Acknowledgments
15
This work was finantially supported by the Iran National
Science Foundation (INSF). This research was funded by the high
impact research (HIR) Grant number: HIR-MOHE-D000001-16001
from the Ministry of Higher Education, Malaysia.
10
References
20
5
0
0
1
2
3
4
Number of PTCAP passes
5
Fig. 5. The dislocation densities for the 1–4 passes PTCAPed specimens calculated
by Eq. (3) based on hn values of Fig. 4.
The dislocation densities were measured by hardness indentation
size effect using the Nix–Gao model. The results showed that
increase in the number of PTCAP passes leads to decrease in the
dislocation densities. The dislocation density is decreased to
2.48 109 cm 2 after fourth pass from 18.1 109 cm 2 after first
pass. TEM results verified calculated values from the Nix–Gao
model. Microhardness of the four passes PTCAP processed tube
increased to 142 HV from initial value of about 62 HV in which
a significant increase had taken place after single pass and it is
saturated in the next passes. There is also superior microstructure
and hardness homogeneity through the processed tube.
[1] R.Z. Valiev, T.G. Langdon, Prog. Mater. Sci. 51 (2006) 881.
[2] R.Z. Valiev, R.K. Islamgaliev, I.V. Alexandrov, Prog. Mater. Sci. 45 (2000) 103.
[3] G. Faraji, H. Jafarzadeh, H.J. Jeong, M.M. Mashhadi, H.S. Kim, Mater. Des. 35
(2012) 251.
[4] V.M. Segal, Mater. Sci. Eng. A 197 (1995) 157.
[5] A.P. Zhilyaev, T.G. Langdon, Prog. Mater. Sci. 53 (2008) 893.
[6] A. Babaei, G. Faraji, M.M. Mashhadi, M. Hamdi, Mater. Sci. Eng. A 558 (2012)
150.
[7] Y. Saito, H. Utsunomiya, N. Tsuji, T. Sakai, Acta Mater. 47 (1999) 579.
[8] M. Barmouz, K. Abrinia, J. khosravi, Mater. Sci. Eng. A 559 (2013) 917–919.
[9] Y.H. Zhao, J.F. Bingert, Y.T. Zhu, X.Z. Liao, R.Z. Valiev, Z. Horita, Appl. Phys. Lett.
92 (2008) 081903.
[10] X. Huang, N. Hansen, N. Tsuji, Science 312 (2006) 249–251.
[11] G. Faraji, M.M. Mashhadi, H.S. Kim, Mater. Lett. 65 (2011) 3009.
[12] G. Faraji, M.M. Mashhadi, H.S. Kim, Mater. Trans. 53 (2012) 8.
[13] G. Faraji, M.M. Mashhadi, K. Abrinia, H.S. Kim, Appl. Phys. A 107 (2012) 819.
[14] G. Faraji, A. Babaei, M.M. Mashhadi, K. Abrinia, Mater. Lett. 77 (2012) 82.
[15] Y. Lin, Y. Zhang, B. Xiong, E.J. Lavernia, Mater. Lett. 82 (2012) 233–236.
[16] S. Graca, R. Colaco, R. Vilar, Surf. Coat. Technol. 202 (2007) 538–548.
[17] X.X. Xu, et al., Mater. Lett. 64 (2010) 524–527.
[18] S. Graca, R. Colaco, P.A. Carvalho, R. Vilar, Mater. Lett. 62 (2008) 3812–3814.
[19] G. Faraji, M.M. Mashhadi, H.S. Kim, Mater. Sci. Eng. A 528 (2011) 4312.
[20] W. Wei, K.X. Wei, G.J. Fan, Acta Mater. 56 (2008) 4771.
[21] A. Habibi, M. Ketabchi, M. Eskandarzadeh, J. Mater. Proc. Technol. 211 (2011)
1085.
[22] A. Mishra, B.K. Kad, F. Gregori, M.A. Meyers, Acta Mater. 55 (2007) 13.
198
G. Faraji et al. / Materials Science & Engineering A 563 (2013) 193–198
[23] A. Kundu, R. Kapoor, R. Tewari, J.K. Chakravartty, Scr. Mater. 58 (2008) 235.
[24] S. Ferrasse, V.M. Segal, K.T. Hartwig, R.E. Goforth, Metall. Mater. Trans. A 28
(1997) 1047.
[25] M.H. Shih, C.Y. Yu, P.W. Kao, C.P. Chang, Scr. Mater. 45 (2001) 793.
[26] F. Dalla Torre, E.V. Pereloma, C.H.J. Davies, Acta Mater. 54 (2006) 1135.
[27] S.J. Kadri, K.T. Hartwig, in: Z. Horita (Eds.), Proceedings of the Nanomaterials
by Severe Plastic Deformation, NanoSPD3, Fukuoka, Japan; 2005,p. 349.
[28] L. Kommel, I. Hussainova, O. Volobuev, Mater. Des. 28 (2007) 2121.
[29] Y. Miyajima, M. Mitsuhara, S. Hata, H. Nakashima, N. Tsuji, Mater. Sci. Eng.
A 528 (2010) 776–779.
[30] Y. Murata, I. Nakaya, M. Morinaga, Mater. Trans. 49 (2008) 20–23.
[31] G. Faraji, P. Asadi, Mater. Sci. Eng. A 528 (2011) 2431–2440.