Professional Documents
Culture Documents
A Model For Stresses, Crack Generation and Fracture Toughness Calculation in Scratched Tin-Coated Steel Surfaces
A Model For Stresses, Crack Generation and Fracture Toughness Calculation in Scratched Tin-Coated Steel Surfaces
b/2) when the crack spacing is smaller than the crack length. For determining the fracture toughness of a 2 m thick
TiN coating on steel substrate a suitable crack eld turned out to be the transversal tensile cracks in the scratched groove. For the studied
case, the fracture toughness of the TiN coating was measured to be K
c
= 7 MPa m
0.5
.
2002 Elsevier Science B.V. All rights reserved.
Keywords: Surface engineering; Coatings; Stress modelling; Fracture; Scratch test; Fracture toughness
1. Introduction
There has been increased interest in the use of coatings
on mechanical components, on tools in the production in-
dustry, on disc drives in the computer industry, on preci-
sion instruments, and on human replacement organs. New
coating deposition techniques developed over the last two
decades offer a wide variety of possibilities to tailor surfaces
with many different materials and structures. In particular,
chemical vapour deposition (CVD) and physical vapour de-
position (PVD) techniques have made it possible to deposit
thin coatings only about 1 m thick in a temperature range
from very high temperatures (about 1000
C) down to room
temperature.
Coating materials such as TiN, TiC, Al
2
O
3
and more re-
cently diamond, diamond-like carbon (DLC) and MoS
2
and
their combinations as multilayers and dopants have been
used with great success. In the best cases, these very thin
an-
gle to the direction of the stylus movement. Second, cracks
forming at the bottom of the groove (type 2: (c) in Fig. 3)
and growing from the groove boundary towards the centre
of the scratch are analysed.
The stress intensity factor is computed following the solu-
tions of Paris [52], Isida [53] and Fichter [54]. The solution
of [52] is the limit of the other ones for crack elds where
the density of the crack eld is high enough to overcome
the effects of crack length. This occurs for the current prob-
lem when crack spacing and crack length are of the same
order. The solutions are given as K =
1
1
is the rst principal stress, b crack spacing and f(a, b) a
non-dimensional function dependent on crack length, a, and
crack spacing/density.
The fracture toughness computation results incorporating
crack density and length interaction effects are presented
K. Holmberg et al. / Wear 254 (2003) 278291 289
Fig. 12. Coating fracture toughness as a function of stylus position in the
groove for angular cracks (type 1) and transversal tensile cracks (type 2)
related to the relation between crack length and crack spacing.
in Fig. 12 as a function of test position for both angular
(type 1) and transverse (type 2) cracks. First of all, it is noted
that different crack types are seen to produce different lev-
els of fracture toughness: angular cracks give values around
4 MPa m
0.5
, while transverse crack fracture toughness is ap-
proximately 7 MPa m
0.5
. The fracture toughness for a similar
TiN coating on steel bent to fracture in a four-point bending
test was calculated based on the experimental data given by
Wiklund et al. [29] and the result was an average K-value
of 7.5 MPa m
0.5
with variation of 1.5 MPa m
0.5
.
The interaction between crack spacing and crack length
is seen to be negligible in comparison to fracture tough-
Fig. 13. Coating fracture toughness as a function of stylus position in the groove for angular cracks (type 1) and transversal tensile cracks (type 2). They
are labelled as mean, lower and upper bounds, which means with respect to rst principal stress values. The stress intensity factor (SIF) is calculated
based on crack density.
ness dependency on stylus position and related scatter. Even
though a wide scale of a/b values is considered, the impact
on fracture toughness is only some percentages and can be
neglected. This means that the analyses can focus solely on
crack density as a characteristic crack eld measure.
By incorporating an estimate of accuracy of nite element
based stress results on characteristic crack scale, the results
of Fig. 13 are attained. They are labelled as lower and upper
bounds, which means with respect to rst principal stress
values. For both crack sizes, especially for transverse type 2
cracks, two effects are noted with respect to the toughness
distribution. Initially, at the beginning of the test and in
particular after about 2.25 mm, the fracture toughness values
begin to rise. For the latter parts this is understood to result
from decoupling of damage, cracking and stress-elds in the
nite element simulations.
At the beginning of the test, the effects of smaller crack
size and its relation to dimensions characteristic of nite el-
ement results are interpreted to increase scatter and uncer-
tainty. In part, this is related to the use of stress values, since
current values are surface values, and initially the coating
through thickness stresses has not stabilised to a pure mem-
brane type of loading. The fact that type 1 angular cracks
produce lower values than type 2 transverse cracks can be
attributed to the analysis procedure, i.e. the applied K solu-
tions are best suited for transversely oriented cracks and do
not account for the orientation exhibited by type 1 cracks.
The solution neglects the associated mixed-mode compo-
nent for the crack driving force and as such produces lower
values of fracture toughness.
This study shows that the scratch test can be used as an
instrument for determining the fracture toughness value of
a coated surface by combining numerical stress simulations
290 K. Holmberg et al. / Wear 254 (2003) 278291
of the coated surface and empirical measurements of the
location and density of the rst cracks to occur. The method
has been protected by patent application [55].
8. Conclusions
A FEM model for the scratch test contact system was
created and applied for the conditions of a 2 m TiN coating
on steel substrate. Based on this the following conclusions
can be drawn.
1. The created model describes the stresses and strains in-
corporating elastic and plastic behaviour in a contact ge-
ometry of the scratch test.
2. In the sliding mode a maximum tensile stress area is
generated in the tail part of the contact area.
3. With increasing load a new stress-eld with a tetra-armed
star shape grows around the contact area. At the tail arms,
stress concentrations are amplied at a distance of 12
times the contact length from the edge of the contact at
the border of the scratch groove. The magnitude of these
stress concentrations is of the same level as within the
contact area.
4. The stress concentrations at the tail arms travel with in-
creasing load to the plane of symmetry and increase in
magnitude considerably faster than within the contact
area, and become the dominating tensile stress as the
contact conditions change to ploughing mode.
5. After about 1 mm of sliding in the scratch test, a peak
area of maximum rst principal stress is generated in
the back-tail region at the border of the scratch groove,
creating the rst visible angular cracks in the coating with
this material-coating combination.
6. In the ploughing mode, the maximum tensile stresses are
located behind the contact at a distance of 0.51 times
the contact length from the back edge of the contact, and
they have a horseshoe shaped ridge of maximum values
with an opening in the sliding direction.
7. The change of the state of deformation from sliding mode
to ploughing mode characterises a loss of load carrying
capacity of the coated surface system.
8. The maximum stresses at the locations of the tetra-armed
stress-eld occur equally throughout the coating, while
the maximum stresses at the points of coating bending
are located just at the top of the coating.
9. The model determines the location of the rst cracks to be
initiated and this information can be used for calculating
the fracture toughness of the coating.
10. Critical fracture toughness is equal to the tensile stress
times the square root of half of the crack spacing (K
c
=