Influence of Surface Effect On Nickel Micro Deep Drawing Process

C. Keller, M. Afteni, M. Banu, A. M. Habraken, E. Hug et al.

Citation: AIP Conf. Proc. 1252, 1025 (2010); doi: 10.1063/1.3457495

View online: http://dx.doi.org/10.1063/1.3457495
View Table of Contents: http://proceedings.aip.org/dbt/dbt.jsp?KEY=APCPCS&Volume=1252&Issue=1
Published by the AIP Publishing LLC.

Additional information on AIP Conf. Proc.

Journal Homepage: http://proceedings.aip.org/
Journal Information: http://proceedings.aip.org/about/about_the_proceedings
Top downloads: http://proceedings.aip.org/dbt/most_downloaded.jsp?KEY=APCPCS
Information for Authors: http://proceedings.aip.org/authors/information_for_authors

Downloaded 25 Sep 2013 to 129.186.1.55. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://proceedings.aip.org/about/rights_permissions
 Influence of Surface Effect On Nickel Micro Deep Drawing
Process
C. Kellera, M. Aftenid, M. Banud, A.M. Habrakena, E. Hugb, S. Castagnec and L.
Duchênea
a
Department ArGEnCo, Division MS²F, University of Liege, Chemin des Chevreuils 1, 4000 Liege, Belgium.
b
Laboratoire CRISMAT, University of Caen, 6 Bd Maréchal Juin, 14050 Caen, France.
c
School of Mechanical and Aerospace Engineering, Nanyang Technological University, 50, Nanyang Avenue,
Singapore 639798.
d
University of Galati, 111 Domneasca St., 800201 Galati, Romania

Abstract. In this paper, the forming behavior of nickel sheets is investigated as a function of the number of grains across
the thickness by finite element simulations. Experimental tensile tests were carried out on nickel samples of
approximately 100 µm grain size and thicknesses ranging between 12.5 µm and 3.2 mm. The decrease of the number of
grains across the thickness involves a decrease of tensile stress due to the apparition of surface effects. These latter were
taken into account for the deep drawing simulation of samples with 250 µm thickness by the use of two different
constitutive elasto-plastic laws for surface and core grains. The simulations with two laws predict a modification of the
distribution of the V.M. equivalent stress and of the damage zones compared to results from simulations using a simple
average law.
Keywords: Nickel, tensile tests, deep drawing, Finite Element simulation.
PACS: 46.50.+a.

INTRODUCTION
As a consequence of the trend of miniaturization on medical, electrical or mechanical devices (Micro Electro
Mechanical Systems, MEMS), the industrial need of micro metal parts has considerably increased since the 1990s.
Thin films or plates of thickness lower than 500 µm are today mass produced by metal processing technologies such
as stamping, blanking or bending. However, due to the miniaturization, size effects may occur and modify both
mechanical behavior and forming ability of these small parts [1-3]. As a consequence, the know-how in metal
forming of massive parts fails to carry out correct forming processes. For the material science community, the
challenge is to understand the physical origin of the behavior modification due to size effects and to design new
solutions for the forming process. For metallic materials, it is now well attested that the most important size effect is
due to the decrease of the number of grains across the thickness [4, 5]. With a decrease of this latter, the flow stress
[4-6], the hardening behavior [5] and the fracture strain [6] are modified. This feature is due to the apparition of
surface effects for specimens with less than a critical number of grains across the thickness (i.e. t/d ratio) which
depends on the grain size and on the stacking fault energy [4]. This apparition of surface effects involves a stress
gradient around 30% between core and surface grains which has been characterized by TEM investigations for
nickel [7].

In Finite Element (FE) simulations of forming process, it is therefore necessary to take into account the
modification of the mechanical behavior of such small samples to predict the forming conditions (force, geometry)
and the appearance of fracture or necking. This point is still under investigation and many authors proposed
constitutive laws that manage to simulate the tensile behavior of small dimension samples in good agreement with
experimental data [8,9] . Nevertheless, many of the constitutive laws proposed by these authors fail to reproduce the
stress distribution into the cross section. This latter can strongly affect the fracture localization and the shape of the
final product.

The objective CREDIT

of this LINE
work (BELOW)
is to propose
TO BE aINSERTED
simple two constitutive
ON THE FIRST PAGElawOF
approach for the simulation of deep
EACH PAPER
EXCEPT
drawing of small nickel samples taking THE PAPER
into account the ON PP. 837
stress TO 840experimentally observed. Nickel is widely
gradient

CP1252, NUMIFORM 2010, Proceedings of the10th International Conference

edited by F. Barlat, Y. H. Moon, and M.G. Lee
© 2010 American Institute of Physics 978-0-7354-0800-5/10/$30.00

1025

Downloaded 25 Sep 2013 to 129.186.1.55. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://proceedings.aip.org/about/rights_permissions
 used in addition with chromium (Ni20%w.t.Cr alloys) in several MEMS due to their thermomechanical and electric
properties. First, experimental tensile tests are carried out to investigate the mechanical behavior of nickel samples
with t/d ratios ranging between 0.1 and 27. Then, the mechanical behavior of core and surface grains are modeled by
two different elasto-plastic constitutive laws. These latter are hence implemented into a finite element code in order
to simulate the micro deep drawing process for samples of 250 µm thickness. Results show that for an identical
punch force, the maximal stress value depends on the single or two constitutive law approaches. For these latter
simulations, the distribution of the damaged zone predicted by the Crockford-Latham criterion is modified.

EXPERIMENTS

Material Description and tensile test results

Polycrystalline 99.99 wt% Ni consisting of various thickness rolled sheets were annealed in vacuum at
temperatures between 600 °C and 1000 °C in order to obtain a mean grain size around 100 µm for each sample.
Thickness values range from 12.5 µm to 3200 µm corresponding to t/d ratios between 0.125 and 27. EBSD analysis
was performed on the unstrained samples of various thicknesses in order to study the grain size. The distribution of
the mean grain diameter d follows a unimodal gaussian statistic. Quantitative texture analysis was also performed to
study the anisotropy of the samples. Pole figures show a very weak crystallographycal texture and calculations
report a coefficient of elastic anisotropy close to 1. The mechanical behaviour is hence fully isotropic.

Dog bone shape samples with a gauge section of 70 mm in length and 12 mm in width were tensile tested at a
strain rate controlled value of 2.4 10-4 s-1. For Ni foils of thickness higher than 250 µm, strain measurements were
achieved by a classical extensometer while for samples with thicknesses lower than 250 µm an optical device was
employed.

Figure 1(a) shows the tensile curves of the nickel samples of different thickness values. Two beams of curves
appear, one for samples with thickness values higher than 500 µm and another one for samples with thicknesses
lower than 125 µm. Inside each beam, all the stress levels are approximately the same independently of the sample
thickness. Figure 1(b) reports the stress level for a strain of 0.1 in tension as a function of the t/d ratio. For samples
with more than four grains across the thickness, the t/d ratio has no influence and the stress level is constant. For
samples with less than four grains across the thickness, a decrease in t/d ratio involves a strong reduction of the
stress level. This critical value for nickel is in good agreement with previous published results [5]. For specimens
with less than one grain across the thickness, the flow stress is not sensitive to the t/d ratio. As previously reported
this stress decrease proceeds from the apparition of a stress gradient between core and surface grains, these latter
being around 10-30% less constrained than core grains [7].

(a) (b)
FIGURE 1. (a) tensile curves of nickel samples with various thicknesses; (b) evolution of the flow stress as a function of the t/d
ratio for a strain level of 0.1.

1026

Downloaded 25 Sep 2013 to 129.186.1.55. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://proceedings.aip.org/about/rights_permissions
 TABLE 1. Results of the fit of the tensile curve by the Voce’s law for t/d=1 and 27
t/d A B C Name
1 17.15 316 7.15 Surface law
27 24 534 6.25 Core law

NUMERICAL SIMULATIONS
The objective of the simulations is to correctly reproduce the size and surface effects that occur for microforming.
One way is to take into account the specific mechanical behaviour of core and surface grains with two constitutive
laws.

Constitutive laws
For the sample with 27 grains cross the thickness, the core grains represent around 92% of the total number of
grains. It could be assumed that the overall mechanical behavior of this sample is due to the core grains. For the
sample with t/d=1, all the grains have two free surfaces. As a consequence, the mechanical behavior of this sample
proceeds from the behavior of the surface grains. Due to the isotropic mechanical behavior and the same grain size
for each sample, the mechanical behavior of core and surface grains can be modeled by a classical J2 Von Mises
elasto-plastic constitutive law with a Voce isotropic hardening [10]. To this aim, the tensile curves of samples with
27 and 1 grains across the thickness were fitted by the equation 1.

σ = A − B(1 − exp(−Cε )) (1)

Table 1 summarizes the values of the three coefficients A, B and C for t/d=1 and 27. These two constitutive laws
were then used to simulate the tensile curves of nickel specimens with various t/d ratios. The simulations were
carried out by the LAGAMINE code of the Liege University [11]. The effect of the grain boundaries were not
directly taken into account due to the simplicity of the elasto-plastic model. The FE mesh is defined in such a way
that three finite elements cover 100 µm, equivalent to the grain size. For surface grains, the two elements below the
free surface are linked to the surface constitutive law. The others elements are linked to the core law. Stress gradient
is then expected to form over 100 µm below the free surface.

Figure 2(a) and 2(b) show respectively the simulated tensile curves and the stress level for a strain of 0.1 for
specimens with various t/d values. The results obtained by the simulations are in very good agreement with the
experimental ones. The two constitutive law approach for the simulation is hence pertinent and can be used to
simulate a micro-deep drawing process.

(a) (b)

FIGURE 2. (a) simulated tensile curves for nickel specimens with different number of grains across the thickness, (b) evolution
of the simulated flow stress as a function of the t/d ratio for a strain level of 0.1;

1027

Downloaded 25 Sep 2013 to 129.186.1.55. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://proceedings.aip.org/about/rights_permissions
 FIGURE 3. Punch force-punch displacement curves of the simulated micro deep drawing process depending on the employed
approach. The sample thickness is 250 µm.

Micro deep drawing simulation

The deep drawing formability was studied for a 250 µm thickness sample with a grain size of 100 µm. This
sample has 2.5 grains across the thickness and is sensitive to the t/d ratio effect described in the experimental section.
In order to be close to experimental conditions, the characteristics of the deep drawing configuration for the
simulations were taken from the experimental work of Saotome et al. [12]. The ratio between the punch diameter
and the sheet thickness was set to 10 while the radius of the matrix and the punch was set to 0.625 mm [12].
Two approaches were used. The first one, named “surface effect approach”, is based on the two constitutive laws
used for the tensile tests. The second approach, named “composite approach”, consists on one single constitutive law,
for core and surface elements, based on a mixture law between core and surface behavior. Simulations were carried
out by the LAGAMINE code with a 0.2 friction coefficient.

Figure 3 shows the punch force-punch displacement curves for the deep drawing process. The two curves are
logically close as the overall behavior of the material was expected to be independent of the approach. Figure 4
shows the Von Mises equivalent stress distribution for the two approaches. At the end of the forming process, the
maximal equivalent stress for the composite approach is approximately 460 MPa and no stress gradient appears
between core and surface regions excepted for the bending zones. For the surface effect approach, stress gradient
developed in the cross section and the maximal equivalent stress is 15% higher than the one for the composite with
531 MPa.

(a) (b)
FIGURE 4. Von Mises equivalent stress distribution in an orthoradial cross-section for Dp/t=10: (a) “surface effect approach” ;
(b) “composite approach”. The sample thickness is 250 µm.

1028

Downloaded 25 Sep 2013 to 129.186.1.55. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://proceedings.aip.org/about/rights_permissions
 (a) (b)
FIGURE 5. Distribution of the Cockroft-Latham damage criterion in an orthoradial cross-section (a) “surface effect approach ;
(b) “composite approach”. The sample thickness is 250 µm.

This increase in maximal equivalent stress may have a strong influence on fracture. In order to study this latter
the Cockroft and Latham fracture criterion was used [13, 14]. Based on the experimental tensile tests, fracture occurs
when the value of the criterion reaches the critical value of 120 J. Figure 5 shows the damage criterion distribution
for the surface effect and the composite approaches. The distribution strongly depends on the employed approach.
For the surface effect one (Fig. 5(a)), the damage zones are spread in the cross section of the blank while for the
composite approach (Fig. 5(b)), they are concentrated on the surface. For the surface effect simulation, the damage
criterion reaches a maximal value of 98 J while the composite approach exhibits a maximal value of 131 J. Fracture
is hence supposed to occur using one constitutive law for the simulations. The forming ability is then underestimated
in this case.

These results show that the two constitutive law approach predicts more accurately the stress distribution
between core and surface regions as the surface effects are taken into account. The damaged zones are strongly
modified which lead to an increase of the predicted formability of the specimen. As a consequence, these
simulations could be used to improve the design of the microforming process, especially those with small tools and
samples. Experimental validation of the fracture criterion for the deep drawing process is in progress.

CONCLUSIONS
The mechanical behavior of nickel is strongly affected by the number of grains across the thickness. With a
decrease of this latter between 1 and 4 approximately, a strong decrease of the flow stress appears due to surface
effects. These surface effects can be correctly taken into account in finite element simulation by a two constitutive
law approach based on the mechanical behavior of surface and core grains. Comparisons between a classical
composite approach and a two constitutive law one show, for a micro deep drawing process, that the distribution of
equivalent stress is greatly modified involving a modification of the damaged zones.

ACKNOWLEDGMENT
The authors acknowledge the Interuniversity Attraction Poles Programme -Belgian State - Belgian Science
Policy (Contract P6/24). The Belgian Fund for Scientific Research FRS-FNRS is also acknowledged.

1029

Downloaded 25 Sep 2013 to 129.186.1.55. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://proceedings.aip.org/about/rights_permissions
 REFERENCES
1. U. Engel, R Eckstein, Journal of Materials Processing Technology 125-126, 35-44 (2002).
2. F. Vollertsen, H. Schulze Niehoff,. Z Hu, International Journal of Machine Tools & Manufacture 46, 1172-1179 (2006)
3. T.A. Kals, R. Eckstein, Journal of Materials Processing Technology,103, 95-101 (2001)
4. S. Miyazaki, K. Shibata, H. Fujita, Acta Metallurgica 27, 855-862 (1979)
5. C. Keller, E. Hug, D. Chateigner, Materials Science and Engineering A500, 207–215 (2009)
6. G. Simons , Ch. Weippert, J. Dual, J. Villain, Materials Science and Engineering A416, 290–299 (2006)
7. C. Keller, E. Hug, X. Feaugas, Mechanics of Materials 42, 44-54 (2010)
8. W. Yun, D. Peilong, X. Zhenying, Y. Hua, W. Jiansping, W. Jingjing, Material and Design 31, 1010-1014 (2010)
9. L. Peng, X. Lai, H.J. Lee, J.H. Song, J. Ni, Materials Science and Engineering A526, 93-93 (2009)
10. M. Tiryakioglu, J.T. Staley, J. Campbell, Journal of Material Science Letter 19, 2179-2181 (2000)
11. A.M. Habraken, M. Bourdouxhe. European Journal of Mechanics A/Solids 11, 381-402 (1992)
12. Y. Saotome, K. Yasuda, H. Kaga, Journal of Materials Processing Technology 13, 641-647 (2001)
13. M.G. Cockroft, D.J. Latham, Journal of the Institute of Metals 96, 33-39 (1968)
14. Y.Y. Zhu, S. Cescotto, A-M. Habraken. Journal of Materials Processing Technology 32, 197-204 (1992)

1030

Downloaded 25 Sep 2013 to 129.186.1.55. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://proceedings.aip.org/about/rights_permissions