Sie sind auf Seite 1von 8

Materials Science and Engineering A361 (2003) 377384

Correlation between nanoindentation and tensile properties


Inuence of the indentation size effect
R. Rodrguez, I. Gutierrez

CEIT and TECNUN, P


o
de Manuel Lardizbal 15, 20018 Donostia-San Sebastian, Basque Country, Spain
Received 24 April 2003; received in revised form 1 July 2003
Abstract
The nanoindentation test has become one of the most broadly expanded techniques used to measure the mechanical properties in a sub-micron
range. However, the interpretation of the data is very difcult due to the Indentation Size Effect (ISE). The ISE can be dened as an increase
of the nanohardness by decreasing the indentation depth of the test. In the present work, the ISE of different metals has been studied by
performing nanoindentation tests at different imposed depths. Furthermore, tensile tests of these materials have been carried out in order to
determine empirical relations between the nanoindentation test results and the tensile properties.
2003 Elsevier B.V. All rights reserved.
Keywords: ISE; Nanoindentation test; Hardness; Tensile tests
1. Introduction
The nanoindentation test is one of the most developed
techniques, which can provide information about the me-
chanical behaviour of the material when it is being deformed
at the sub-micron scale. The method developed by Oliver
and Pharr [1] allows determining the hardness and the elastic
modulus from the nanoindentation load-displacement data.
However, it has been observed by several authors that hard-
ness in nanoindentation tends to be overestimated due to
a high size-dependence [2,3] of the tests. This size depen-
dence observed in nanoindentation is called the Indentation
Size Effect (ISE).
Generally speaking, ISE means that hardness at small
depth is much greater than at greater depths [4]. This ef-
fect becomes pronounced when the indentation is shallower
than 10 m. Therefore, on the nanoindentation range, the
hardness value will be strongly dependent on the depth of
the indentation. This is the reason why the interpretation of
the nanohardness measurements is difcult. Nevertheless,
the ISE has not only been observed in nanoindentation tests
but it has also been detected for low load Vickers measure-
ments [5].

Corresponding author. Tel.: +34-943-212800; fax: +34-943-213076.


E-mail address: igutierrez@ceit.es (I. Gutierrez).
This effect on the indentation measurements cannot be
explained by using a continuum mechanics based analysis
because, according to this theory, the hardness value should
be independent of the applied load, or equivalently, of the
indent size [6]. Several mechanisms to explain the ISE have
been suggested, including the effect of the friction between
the indenter and the sample [7], the presence of a work hard-
ened surface layer [8], the lack of measurement capabilities
at shallow indentations or the presence of surface layers,
oxides and chemical contamination [9].
However, this is not the only size-dependent phenomenon
related to the material mechanical behaviour. Several exam-
ples have been found in the literature such as the HallPetch
grain-size effect [10], the higher strengthening in metals by
a given volume fraction of hard particles when the parti-
cles are small [11] or, in torsion, the thinner is the wire the
greater is the strength of the material [12].
In order to explain these size-dependences it was sup-
posed by different authors that the strengthening effect was
intrinsically linked to non-uniform deformation. By using
this concept, the strain gradient plasticity (SGP) model was
developed. This model includes a material length scale by
using the concept of geometrically necessary dislocations as
a function of the strain gradients. Ashby [13], who made
the distinction between two kinds of dislocation, described
this concept as follows. The density of dislocations that
are accumulated randomly are known as statistically stored
0921-5093/$ see front matter 2003 Elsevier B.V. All rights reserved.
doi:10.1016/S0921-5093(03)00563-X
378 R. Rodrguez, I. Gutierrez / Materials Science and Engineering A361 (2003) 377384
dislocations (
S
). This density of dislocations is a function
of the total uniform strain of the sample and, with pyrami-
dal indenters, is supposed not to be size-dependent. On the
other hand, the dislocations that result from the gradients of
plastic shear strain are known as the geometrically necessary
dislocations (
G
). The magnitude of
G
is given by:


b
, (1)
where is a constant, the macroscopic plastic shear strain,
b the Burgers vector and the local length scale of the
deformation eld related to the indentation size.
In this way, the total density of dislocations (
T
) can be
dened as the sum of both contributions:

T
=
G
+
S
. (2)
The strain hardening of the materials is caused by dis-
locations interacting with each other, and with barriers that
impede their motion through the crystal lattice. Taylors re-
lation denes the contribution of dislocations to the ow
stress:
= Mb

T
= Mb

(
S
+
G
), (3)
with being a constant, M the Taylor factor and the elastic
shear modulus. At very large indentation sizes, it can be
assumed that
S
is dominant and the inuence of
G
is very
small. However, the smaller the length scale of the gradient,
the more important the effect becomes. Consequently, in a
submicron range of indentation, the contribution of the strain
gradient effect can be comparable to the hardening induced
by
S
.
In the present work, the depth dependence of the hardness
of different materials has been studied. A comparison has
been made between the nanohardness measured at different
depths and the tensile properties of the samples. Finally, by
using the concept of geometrically necessary dislocations,
the relationship between the hardness due to statistically
stored dislocations alone and the tensile properties of the
materials has been studied.
2. Experimental procedure
Several steels were investigated to span a wide range
of mechanical properties. They include different ferritic,
Table 1
Ferrite grain size and chemical composition of the steels employed in the present work
Sample Phase C (%) Mn (%) Si (%) S (%) P (%) V (%) Nb (%) Grain size (m)
M1 Martensite 0.42 0.68 0.18 0.24 0.009
M2 Martensite 0.29 1.5 0.26 0.004 0.011
M3 Martensite 0.16 1.5 0.25 0.004 0.012
P1 Pearlite 0.82 0.73 0.25 0.006 0.008
F1 Ferrite 0.058 0.45 0.016 0.004 0.012 0.02 5.5
F2 Ferrite 0.064 0.70 0.008 0.004 0.01 0.042 0.028 4.4
F3 Ferrite 0.06 1.085 0.014 0.006 0.015 0.11 0.029 3.4
pearlitic and martensitic steels. Details concerning the mi-
crostructure and the chemical composition are given in
Table 1. Moreover, a 70Cu30Zn brass processed under
different conditions, in order to produce two different grain
sizes, has been tested. The samples identied as B
1
and
B
2
have mean grain sizes of 115 and 60 m, respectively.
Additionally, a 1% Mn cold rolled aluminum alloy, A1, was
tested.
The preparation of the surface prior to the nanoindenta-
tion measurements has been carried out as follows. The me-
chanical polishing of the samples consisted in rst grinding
the specimens with SiC papers of decreasing grid size and
then polishing them with diamond pastes of 6, 3 and 1 m.
Finally, they were polished very slowly with colloidal sil-
ica suspension of 0.3 m, for approximately an hour. The
aim of the silica polishing is not only to minimise the strain
hardening on the sample surface but also to effect a slight
etching that allows the observation of the microstructure by
SEM.
All experiments were performed at different imposed
depth (50, 113, 250, 340, 650, 900, 1500, 2000 and
3000 nm) using a diamond Berkovich indenter tip. An
experiment at the maximum load of the nanoindenter
(620 mN) has also been performed. The load-displacement
curves were analysed by the method proposed by Oliver and
Pharr [1].
Arrays of at least 40 indentations for each imposed
depth were performed for all samples. At low indenta-
tion depths, the roughness on the interface between two
grains or different microstructural components does not
allow a correct interpretation of the data. Therefore nanoin-
dentations carried out on these zones were not taken into
account. Conventional etching by using a nital solution
has been used to reveal the microstructure of the differ-
ent steel samples. The ferrite grain size and the pearlite
interlamellar spacing were determined by quantitative met-
allography. For the brass alloy a FeCl
3
solution in ethanol
was used to reveal the grain size, while for the aluminium
alloy a Barkers reagent has been applied electrolytically
[14].
The tensile tests have been performed on a 4505 Instron
testing machine under strain control. The tests have been
carried out at a strain rate of 10
3
s
1
. The 0.2% proof stress
and the true stress at fracture were determined from recorded
load-elongation curves.
R. Rodrguez, I. Gutierrez / Materials Science and Engineering A361 (2003) 377384 379
Fig. 1. Micrographs of some of the samples used in the present work: (a) martensitic steel, (b) pearlitic steel, (c) ferritic steel, (d) brass alloy and (e)
aluminium.
3. Results and discussion
3.1. Microstructures
The most characteristic microstructures of the materials
used in the present work are shown in Fig. 1. The rst mi-
crograph corresponds to a SEM image of the martensite in
Steel M1. The second example is the pearlitic microstruc-
ture with an interlamellar spacing, = 0.17 m in sample
P1. The micrograph in Fig. 1c shows the essentially ferritic
microstructure in the F1 low carbon steel and nally the two
last examples correspond respectively to the B1 -brass and
to the cold rolled AlMn alloy.
3.2. Tensile tests
The tensile curves of the different materials are plotted in
Fig. 2. As expected, a broad range of strengths is obtained.
The softest materials are the Al-alloy and the B1 -brass.
At the other end of the spectrum, the strongest materials
correspond to the martensitic steels. The strength of such
microstructures is directly related to the carbon content of
380 R. Rodrguez, I. Gutierrez / Materials Science and Engineering A361 (2003) 377384
Aluminum
0
300
600
900
1200
1500
1800
2100
0 0.1 0.2 0.3 0.4 0.5

(
M
P
a
)
Martensitic
Pearlitic Steel
Ferritic Steels
Brass alloys

Fig. 2. Tensile curves.


the steel. Both, the yield stress and tensile strength increase
with the carbon content of the sample, which is the usual
behaviour reported by different authors [15,16]. The yield
stress of martensitic steels used in the present work ranges
from 1052 to 1632 MPa (Table 2). The tensile strength corre-
sponding to the high carbon martensite in steel M1 (0.42%C)
is not available due to an early cracking of the sample.
The yield stress of the pearlitic steel reaches 670 MPa,
while for ferritic steels the value ranges from about 400
to 600 MPa, depending on the steel. In these samples, the
value of the yield strength depends strongly on the chemical
composition and on the grain size [17]. The higher the alloy
content and the smaller is the grain size, the stronger is the
material. A HallPetch relation is generally found to express
the contribution,
g
, of the grain size, D, to the yield stress
[18]:

g

= 18D
1/2
. (4)
In the present case, the dependency of the experimental
yield stress on the grain size is stronger than that predicted
by this expression, which is an indication of some precipita-
tion hardening contributing to the tensile behaviour of these
steels.
In the -brass alloy, the grain size also plays an important
role. When the grain size decreases from 115 to 60 m, the
Table 2
Tensile properties of the materials used in the present work
Sample Phase
y
(MPa)
u
(MPa) Strain hardening
exponent (n)
M1 Martensite 1632
M2 Martensite 1338 2014 0.20
M3 Martensite 1025 1457 0.155
P1 Pearlite 670 1192 0.23
F1 Ferrite 389 511 0.16
F2 Ferrite 478 622 0.15
F3 Ferrite 609 723 0.14
B1 -Brass 93 452 0.18
B2 -Brass 240 474 0.18
A1 Aluminium 133 203 0.13
Fig. 3. Nanoindentation array in a martensitic steel for an indentation
depth of 250 nm.
value of the yield strength increases strongly from 94 to
239 MPa. However, the tensile strength of both brass samples
is quite similar (452 and 474 MPa). Finally, the aluminium
alloy, with a yield stress of 133 MPa and a tensile strength
close to 200 MPa, is the softest of all the materials in the
present work.
3.3. Nanoindentation tests
Nanoindentation tests have been performed on all the sam-
ples. An example of the array of indentations carried out
on each sample is shown in Fig. 3, for the martensitic mi-
crostructure in steel M1. Additionally, two examples of the
typical load-depth curves are shown in Fig. 4, for the softest
and one of the strongest of the investigated materials.
At each imposed depth, the true indentation elastic mod-
ulus, E, and the nanohardness, H, can be deduced from the
curves [1]. In Fig. 5, the mean value of the elastic modulus is
plotted versus the depth of the indentation. For aluminium,
an almost constant value for E is obtained over the different
applied test conditions, which seems reasonable. A value of
80 GPa is obtained which is in fairly good agreement with
the reported Youngs modulus in literature (70 GPa) [19].
The same is true for brass that leads to a value of 120 GPa,
relatively close to the 110 GPa reported [19]. In steel, the
discrepancies are larger between experiment and the
Aluminum
Martensite
(0.29%C)
0
100
200
300
400
500
0 500 1000 1500 2000 2500
Displacement(nm)
L
o
a
d
(
m
N
)
Fig. 4. Nanoindentaion load-depth curves for aluminium and a martensitic
steel.
R. Rodrguez, I. Gutierrez / Materials Science and Engineering A361 (2003) 377384 381
Martensite
Ferrite
Brass
Aluminum
Pearlite
0
100
200
300
0 1000 2000 3000
Depth(nm)
E
l
a
s
t
i
c

M
o
d
u
l
u
s
(
G
P
a
)

Fig. 5. Mean Elastic Modulus as a function of the indentation depth.
common value deduced from the literature (200 GPa) [19].
The deviation fromthese standard values takes place because
the indentation elastic modulus values are calculated by us-
ing the contact area from the indentation depth considered
since the rst contact point with the initial at surface, as
described by the Oliver and Pharr method [1]. However, the
plastically deformed zone around the indentation can pile-up
against the indenter or sink-in, depending on the work hard-
ening of the material. The consequence of this behaviour is a
slight deviation fromthe standard values of the elastic modu-
lus [20]. For the case of pearlite, the discrepancies are larger
and may be due to a higher degree of complexity due to the
presence of two phases with different strengths. Texture is
another source of deviation of the indentation modulus from
Youngs modulus. For brass, Vlassak and Nix [21] determine
25% of scatter between indentation modulus determined
for different crystallographic orientations, but even higher
discrepancies can be found, depending on the material.
The graph in Fig. 6 shows the mean value of the nanohard-
ness measurements as a function of the imposed depth. The
shape of the curves is similar for all samples, except the one
corresponding to the high carbon martensitic steel, M1. In
general, the nanohardness is slightly dependent on the im-
posed depth when the depth exceeds 1 m. However, as the
Pearlitic steel
aluminum
0
1
2
3
4
5
6
7
8
9
10
11
0 2000 4000 6000 8000
h(nm)
H
(
G
P
a
)
M1 M2
M3 P1
F3 F2
F1 B2
B1 A1
Martensitic steels
Ferritic steels
Brass
ll
Fig. 6. Hardness as a function of the depth for different materials.
indentation depth decreases below 1 m, a rapid increase
of the nanohardness value is observed. For example, the
hardness for the F1 ferritic steel increases from about 1.85
to 3.3 GPa when the indent depth decreases from 1500 to
50 nm. In martensitic steels, the ratio of the nanohardness
at 50 nm to the nanohardness at 900 nm is about 1.5, and in
aluminium and brass alloys this ratio is greater than 2.
For the M1 martensitic steel, a continuous reduction of
the nanohardness with increasing depth is observed over the
entire range of indentation depths. As a consequence, the
curve corresponding to this steel crosses, at a certain depth,
those obtained for softer materials. As an example, for a
depth of 1 m the value of the nanohardness of the 0.42%C
martensite becomes lower than those obtained for the 0.16
and 0.29%C martensites. The 0.42%C sample has been ob-
served by SEM in order to nd an explanation to this rapid
decrease of the nanohardness value. However, no apparent
microcracks, forming around the high depth nanoindenta-
tion tests, has been detected. This irregular behaviour may
be due to the effect of residual stress that remain in the sam-
ple after quenching.
The nanohardness data in Fig. 4 were calculated using
the Oliver and Pharr method [1] that does not take into
account elasticplastic effects. Gken et al. [22] proposed
a method correcting pile-up effects and possible surface
roughness. This method allows determining a correction fac-
tor for nanohardness based on the relation between the in-
dentation modulus and Youngs modulus. This method has
been applied to the results in the present work and was dis-
carded due to the increase of the nanohardness scatter it
produces for a given depth. This effect is probably due to
the polycrystalline nature of the present samples and the
dissimilar effect anisotropy has on the indentation modulus
and on nanohardness. Vlassak and Nix [21] reported that E
is highly dependent on the crystallographic orientation but
that hardness is not.
A linear relation is usually reported to relate microhard-
ness to the tensile properties of the materials [23]. A similar
type of comparison has been carried out between nanohard-
ness and both the yield stress and the tensile strength. The
graph in Fig. 7 shows these relations at different imposed
depths for the materials in the present work. As a rst ap-
proximation, a linear relationship can be drawn to t each set
of experimental data. The point corresponding to a nanoin-
dentation carried out at 50 nm depth in pearlite clearly de-
viates from the general trend. This is probably due to a
nanohardness indentation being lower than the interlamel-
lar spacing. The other point to deviate signicantly from the
straight line is the one corresponding to the -brass alloy
with a large mean grain size (115 m). This point deviates
from the general trend for all the considered indentation
depths. Probably due to its large grain size, the indentation,
which is a local measure, gives lower values than expected
based on the macroscopic yield stress.
Fig. 7 shows that the slopes of the tting lines de-
crease slightly by increasing the depth of the indentations.
382 R. Rodrguez, I. Gutierrez / Materials Science and Engineering A361 (2003) 377384
Fig. 7. Nanohardness vs. (a) yield stress and (b) tensile strength at different
imposed depths.
Furthermore, the intercept point of the linear tting ap-
proaches to zero as the depth of the indentation increases.
However, it can be assumed that, for large indentation depth
tests, Tabors expression [23] can be used to relate the
nanohardness to tensile properties:
H c. (5)
For the experimental points corresponding to a depth of
2 m, the value of c is about 4.3 and 2.8 when considering
the yield stress and the tensile strength respectively, which
are relatively close to the value of 3 generally accepted [23].
For lower indentation depths, in agreement with the results
in Fig. 7, the above relation needs to be modied in the
following way:
H = b +c

, (6)
with b and c

depending on the imposed depth. As an exam-


ple, for an imposed depth of 50 nm the value of c

in equa-
tion relating H to
u
is close to 3.85, while b approaches
937 MPa.
In order to explain the Indentation Size Effect, Nix and
Gao [3] proposed the following relationship by using the
concept of geometrically necessary dislocations for nanoin-
Fig. 8. Square of the nanohardness value against the inverse of the depth:
(a) different steel microstructures and (b) aluminium and brass alloys.
dentation test:
H
H
0
=

1 +

. (7)
H is the nominal hardness for a given depth; h, and h

are
characteristic depths which depend on the shape of the in-
denter and on the material. Finally, H
0
can be dened as the
hardness that would arise from the statistically stored dislo-
cations alone, or equivalently, the hardness obtained in the
limit for an innite depth.
In Fig. 8, the square of the nanohardness value obtained in
the indentation tests is plotted as a function of the reciprocal
of the indentation depth. It can be seen that a linear relation is
closely followed for the most of the materials, in agreement
with Eq. (7). The value of the hardness at innite depth,
H
0
, can be estimated by extrapolating the mentioned linear
relations to (1/h) = 0. The obtained values are shown in
Table 3.
The value of H
0
can be dened as the hardness due to

s
alone. When the H
0
values are plotted against the yield
stress or the tensile strength for each sample, see Fig. 9,
good linear ts are obtained, leading to Tabor relations of
the form:
H
0
= 4.15
y
, (8)
H
0
= 2.78
u
. (9)
R. Rodrguez, I. Gutierrez / Materials Science and Engineering A361 (2003) 377384 383
Table 3
Values of h

and H
0
obtained in the present work
Sample h

(m) H
0
(GPa) (GPa) b (m)

M1 0.21 6.7 81 2.58 10


4
0.60
M2 0.23 5.6 81 2.58 10
4
0.53
M3 0.27 3.9 81 2.58 10
4
0.4
P1 0.33 2.9 81 2.58 10
4
0.32
F1 0.20 2.5 81 2.58 10
4
0.23
F2 0.40 1.8 81 2.58 10
4
0.23
F3 0.35 1.6 81 2.58 10
4
0.21
B1 0.80 0.9 37 2.50 10
4
0.35
B2 0.28 1.4 37 2.50 10
4
0.33
A1 0.38 0.5 25.4 2.86 10
4
0.19
The value of

has been deduced by using the indicated values of


and b and M = 3.
The values for the slope are relatively close to the com-
monly accepted value of 3 for the Tabors factor and are
in close agreement with the values deduced before for the
slope of the lines in Fig. 7 corresponding to an indentation
depth of 2 m.
The plots in Fig. 8 allow, additionally, to deduce the value
of h

for each material from the slope of the straight lines.


The obtained values are shown in Table 3. The slope, s, can
be expressed as:
s = h

H
2
0
=
27
2
M
2
b
2

2
tan
2
, (10)
with as the angle between the surface of the indenter and
the plane of the surface being indented. From this equa-
tion, and taking into account the shear modulus values for
steel (81 GPa), brass (37 GPa) and aluminium (25 GPa) and
M = 3, the value of in Eq. (10) can be estimated. The ob-
tained values are shown in Table 3. The value of calculated
for the aluminium alloy is slightly lower than the value of
0.24 obtained in [24]. For brass, a mean value of = 0.34
has been obtained which is relatively close to that used by
GilSevillano for copper [25] and agrees well with the value
= (0.5/

3) = 0.29 obtained in [3] from nanoindenta-


y = 4.15x
R
2
= 0.99
y = 2.78x
R
2
= 0.96
0
2000
4000
6000
8000
0 500 1000 1500 2000 2500
(MPa)
H
0
(
M
P
a
)
yield stress
tensile strength
Fig. 9. Values of H
0
obtained by linear tting against the yield stress and
the tensile strength of the sample.
y = 0.0304x
R
2
= 0.9781
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0 5 10 15 20
(
u
-
y
)/
t
[GPa]

2
b
[
G
P
a
2

m
]
Fig. 10. Plot of the product (

)
2

2
b as a function of the strain hardening
of the sample.
tion tests in brass. For ferrite, the values of (in Table 3)
are slightly lower than those deduced from torsion [25] or
tensile [26] tests. However, the values are within the vari-
ation range generally found in the literature. As reviewed
in [27], for single and polycrystalline iron, the values of
the product M range from 0.7 to 1.4. Assuming for M a
value of 3 and taking the values obtained in the present
work for steel, approximately the same range of variation is
reached. It is to be noted that, instead of a constant that
would be expected, a variation of this parameter with the
strength of the material agrees better with the experimental
results.
The dependence on the elasticplastic behaviour of the
material is related to the response of the material being in-
dented [28]. As has been mentioned above, several authors
have suggested that the degree of sink-in or pile up of the
materials can be expressed as a function of the work harden-
ing exponent [29,30]. During the indentation, materials with
a low work hardening exponent accommodate the volume of
material ejecting it to the sides of the indenter (pile-up). In
the same way, in materials with a high value of n (n > 0.3)
the sink-in effect takes place. In both cases, the contact area
is different from the cross-sectional area estimated by the
method described by Oliver and Pharr [1]. Consequently,
there is a deviation between the real and the computed area
that is controlled by the elasticplastic behaviour. This de-
viation can affect in an articial way the values calculated
by using Eq. (10).
In order to differentiate between in Eq. (3) and its appar-
ent value obtained from the ts carried out using Eq. (10),
now

is used to denote this last, leading to:

= f, (11)
with f being a geometrical correction factor to take into
account the correct contact area. In Fig. 10, the value of

2
b is plotted as a function of (
u

y
)/
t
. The difference
between the tensile strength and the yield stress divided
by the total strain is a measure of the strain hardening of
384 R. Rodrguez, I. Gutierrez / Materials Science and Engineering A361 (2003) 377384
the materials. A linear relationship can be deduced of the
form:

2
b = (3 10
2
)

t
(12)
to t reasonably well to the experimental results.
4. Conclusions
A linear relationship can be found between the nanohard-
ness and both the yield stress and the tensile stress of the
material. However, due to the indentation size effect, the
slope and the intercept of such a relation depend on the in-
dentation depth.
For indentation depths higher than about 2 m, Tabors
relation applies with coefcients of 4.3 and 2.8 for the yield
stress and the tensile strength, respectively. These values are
not far from the value of 3 generally found for the tensile
strength.
The model proposed by Gao and Nix, which takes into
account the geometrically necessary dislocations, agrees
reasonably well with the experimental results. However,
a correction factor is required in order to describe prop-
erly the true interaction between the indenter and the
sample.
Acknowledgements
Part of this work has been founded by the European Com-
mission (ECSC Project no. 7210-PR-166) and the CICYT
Spain (Project no. MAT2000-1803-CE).
References
[1] W.C. Oliver, G.M. Pharr, J. Mater. Res. 7 (1992) 1564.
[2] J.W. Hutchinson, Int. J. Solids and Struct. 37 (2000) 225.
[3] W.D. Nix, H. Gao, J. Mech. Phys. Solids 46 (1998) 411.
[4] N.I. Tymiak, D.E. Kramer, D.F. Bahr, T.J. Wyrobek, W.W. Gerberich,
Acta Mater. 49 (2001) 1021.
[5] K. Miyahara, S. Matsuoka, T. Hayshi, Metall. Mater. Trans. 32A
(2001) 761.
[6] A.A. Elmustafa, D.S. Stone, J. Mech. Phys. Solids 51 (2003) 357.
[7] H. Li, A. Chosh, R.C. Han, Bradt, J. Mater. Res. 8 (1993) 1028.
[8] H. Pelletier, J. Krier, A. Cornet, P. Mille, Thin Solid Films (2000)
147.
[9] M. Atkinson, J. Mech. Sci. 33 (1991) 843.
[10] N.J. Petch, J. Iron Steel Inst. 174 (1953).
[11] H.H.M. Cleveringa, E. Van der Giessen, A. Needleman, Acta Mater.
45 (1997) 3163.
[12] N.A. Fleck, G.M. Muller, M.F. Ashby, J.W. Hutchinson, Acta Metal.
Mater. 42 (1994) 475.
[13] M.F. Ashby, Phil. Mag. 21 (1970) 399.
[14] J.R. Davies & Associates (Eds.), ASM Speciality Handbook, Alu-
minium and Aluminium Alloys, ASM International, Materials Park,
OH, 1993, p. 488.
[15] M. Cohen, Trans. AIME 224 (1962) 638.
[16] W.C. Leslie, Sober, Trans. ASM 60 (1967) 459.
[17] F.B. Pickering, Cahn et al. (Eds.), Materials Science and Technology,
vol. 7, 1993, Chapter 2, pp. 4594.
[18] T. Gladman, The Physical Metallurgy of Microalloyed Steels, The
Institute of Materials, London, 1997, p. 39.
[19] W.D. Callister, Materials Science and Engineering, Wiley, New York,
1990, pp. 738739.
[20] J. Gil Sevillano, P. Buessler, J. Vrieze, W. Kaluza, O. Bouaziz, Th.
Iung, E. Bonifaz, A. Martin Meizoso, J.M. Martinez Esnaola, I.
Ocaa, ECSC Steel RTD Final report, CECA7210-PR-044 (2000).
[21] J.J. Vlassak, W.D. Nix, J. Mech. Phys. Solids 42 (1994) 1223.
[22] M. Gken, M. Kempf, W.D. Nix, Acta Mater. 49 (2001) 903.
[23] D. Tabor, Hardness of Metals, Clarendon Press, Oxford, 1951.
[24] N. Hansen, X. Huang, Acta Mater. 46 (1998) 1827.
[25] J. Gil Sevillano, J. Phys. III (1991) 967.
[26] R. Rodriguez, I. Gutierrez, Unied Formulation to Predict the Tensile
Curves of Steels with Different Microstructures, Proc. Thermec
2003: International Conference on Processing and Manufacturing of
Advanced Materials, Materials Science Forum, Vol. 426432 (2003)
4525.
[27] M. Dollar, I.M. Bernstein, A.W. Thompson, Acta Metall. 36 (1988)
311.
[28] K.W. McElhaney, J.J. Vlassak, W.D. Nix, J. Mater. Res. 13 (5)
(1998) 1300.
[29] A.L. Norbury, T. Samuel, J. Iron Steel Inst. 117 (1928) 673.
[30] F.A. McClintock, S.S. Rhee, in: Proceedings of the 4th US National
Conference, Applied Mechanics 2, ASME, Berkeley, CA, 1962,
p. 1007.

Das könnte Ihnen auch gefallen