...

Friedrich2011-PolymerScratching.pdf

by user

on
Category: Documents
9

views

Report

Comments

Transcript

Friedrich2011-PolymerScratching.pdf
Tribology International 44 (2011) 1032–1046
Contents lists available at ScienceDirect
Tribology International
journal homepage: www.elsevier.com/locate/triboint
Scratch resistance of high performance polymers
K. Friedrich a,b,n, H.J. Sue c, P. Liu c, A.A. Almajid b
a
b
c
Institute for Composite Materials (IVW GmbH), Technical University of Kaiserslautern, 67663 Kaiserslautern, Germany
CEREM, College of Engineering, King Saud University, Riyadh, Saudi Arabia
Polymer Technology Center, Texas A&M University, College Station, Texas, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 11 January 2011
Received in revised form
4 April 2011
Accepted 11 April 2011
Available online 20 April 2011
Scratch tests were carried out on various high performance polymers, including (1) polybenzimidazole
(PBI), (2) polyparaphenylene (PPP), (3) polyetheretherketone (PEEK), and (4) polyimide (PI). The scratch
damage features were characterized using laser confocal and scanning electron microscope. Scratch
resistance at room temperature decreased in the same order as the materials are listed above. It was
attempted to correlate the scratch depth with basic mechanical properties, such as Young’s modulus,
tensile strength, and scratch hardness. Also, the scratch coefficient of friction was considered as a
possible measure to differentiate between the various materials tested.
& 2011 Elsevier Ltd. All rights reserved.
Keywords:
Scratch
High performance polymers
Elastic modulus
1. Introduction
The scratch process can be defined as a mechanical deformation process where a controlled force or displacement is exerted
on a hard tip to indent onto another material substrate and
move across its surface at a prescribed speed [1]. In terms of
wear as a sub-discipline of tribology, scratch can be considered
as a single path, single asperity wears process with high
abrasion characteristics (Fig. 1). The typical mechanisms of
abrasive action can be categorized as microplowing, microcutting, and microcracking [2].
The scratch performance of polymers has caught significant
attention in the past few years because of their greatly expanded
usage in the electronic, optical, household, and automotive
applications, where long term esthetics is important. Unlike
ceramics and metals, polymers are particularly susceptible to
visible surface deformation and damage, even under low contact
loads. Various scratch induced damage features, such as mars,
fish-scales, parabolic cracks, and material removals have been
observed from a wide variety of polymeric materials [3–6].
Majority of these observations have been made, however, only
for commodity or engineering polymers, such as polyethylene,
polypropylene, polystyrene, polycarbonate, epoxy, etc.; whereas,
the investigation of the scratch behavior of existing or newly
developed high performance polymers is still lacking. It was
therefore the objective of this study to carry out a series of
n
Corresponding author at: Institute for Composite Materials (IVW GmbH),
Technical University of Kaiserslautern, 67663 Kaiserslautern, W-67653
Kaiserlautern, Germany.
E-mail address: [email protected] (K. Friedrich).
0301-679X/$ - see front matter & 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.triboint.2011.04.008
scratch tests on various high performance polymers, of which
some of them have already been used for other tribological
applications, such as bearings or sliding elements.
2. Experimental details
2.1. Materials
The polymers chosen for the present scratch tests are listed in
Table 1. All materials, except PEEK, were received from the
manufacturers as compression molded or extruded plates. PEEK
was prepared in-house in the form of 3 mm thick dog bone
samples using an Arburg injection molding machine. All the
material properties listed in Table 1 were taken from data sheets
published by the manufacturers [7–10]. Regarding the special
characteristics of the polymers chosen, the following short statements can be made:
(a) Polybenzimidazole (PBI)
PBI is an imidized thermoplastic, being the highest performing engineering plastic currently available. It offers the highest heat resistance (heat deflection temperature of 427 1C)
and mechanical property retention over 205 1C of any unfilled
plastic. Also, its wear resistance and load carrying capabilities
at extreme temperatures is better than that of many other
reinforced or unreinforced engineering plastics. PBI is also an
excellent thermal insulator. Other plastics in melt do not stick
to PBI. These characteristics make it ideal for contact seals
and insulator bushings in plastic production and molding
equipment [7].
K. Friedrich et al. / Tribology International 44 (2011) 1032–1046
1033
Tribology
Lubrication
Wear
Friction
(Surface Damage with /without Remarkable Material Removal)
Abrasion
Adhesion
(SharpTips Scratching)
(Sliding against Smooth Surface)
Surface Fatigue
Tribo-Corrosion
(Rolling Counterpart)
(Chemical Interactions)
Single Pass
Multiple Pass
(Repeatedly against
the Same Counterpart)
Many Asperities
Single Asperity
(e.g. Sand Paper)
(e.g. Tip of Nail)
Reduction of
Abrasiveness
with Time
Many Scratches
Simultaneously
Single Scratch
(Material Sees always
a Fresh Counterpart)
Micro-Plowing
Micro-Cutting
Micro - Cracking
(e.g. Mars)
(e.g. Chips)
(e.g. Fish Scale Pattern)
(w/wo Material Removal)
Fig. 1. Schematic relationship between the terms tribology, wear, abrasion, and scratch.
Table 1
List of materials investigated, including manufacturer and important properties (E¼ elastic modulus; sB ¼strength; anI ¼notched Izod impact energy; eB ¼strain at break;
and r ¼density).
anl (J/m)
r (g/cm3)
3
28
1.301
207
5
43
1.210
3.55
2.95
100
80
34
7.6
54
46
1.301
1.401
2.76
61
8.5
53
1.427
Short name
Manufacturer company code
E (GPa)
sB (MPa)
PBI
PBI performance prod.
Celazole U60 SD
Solvay adv. polymers
PrimoSpire PR 120
Victrex PEEK 450 G
DuPont (Vespel)
ST 2002 (2% Gr)
DuPont (Vespel)
ST 2010 (10% Gr)
5.86
159
8.30
PPP
PEEK
PI 2
PI 10
(b) Polyparaphenylene (PPP)
PPP can be considered as a self-reinforced polymer, which at
room temperature is one of the stiffest and strongest unreinforced plastics to-date. Even without fiber reinforcement, PPP
delivers tensile properties that are comparable to those of
many reinforced plastics. Additional benefits are their lighter
weight and no loss of ductility. These properties combine
with high compressive strength—one of the highest among
plastics—makes PPP an excellent candidate for weight-sensitive applications that have historically relied on composites
and specialty metals for superior mechanical performance [8].
(c) Polyetheretherketone (PEEK)
PEEK is a semicrystalline thermoplastic with excellent
mechanical and chemical resistance properties that are
retained at high temperatures. It exhibits Young’s modulus
eB (%)
of 3.6 GPa and tensile strength of 90–100 MPa. PEEK has a
glass transition temperature at around 143 1C and melts at
around 343 1C. It is highly resistant to thermal degradation as
well as attacked by both organic and aqueous environments.
Because of its robustness, PEEK is used to fabricate items used
in demanding applications, including bearings, piston parts,
pumps, compressor plate valves, and cable insulation. It is one
of the few plastics suitable for ultra-high vacuum applications. PEEK is considered an advanced biomaterial used in
medical implants. It is extensively used in the aerospace,
automotive, electronic, and chemical process industries [9].
(d) Polyimide (PI)
The chemical structure of a typical polyimide is called
aromatic, as all of the carbons in the polymer chain are part
of either the imide or the benzene (aromatic) rings. This
1034
K. Friedrich et al. / Tribology International 44 (2011) 1032–1046
arrangement leads to a higher glass transition temperature
(if one can be detected) and greater thermal and oxidative
stability than the majority of organic polymers. Thermosetting polyimides are also known for their good chemical
resistance and excellent mechanical properties. They exhibit
very low creep and high tensile strength. The polyimide
materials are lightweight, flexible, and resistant to heat and
chemicals. Therefore, they are used in the industry for
flexible cables, as insulating films, as bearings, and for
medical tubing [10].
2.2. Hardness testing
The Martens hardness HM (see Ref. [11]) of different materials
was determined by the use of a Dynamic Ultra Micro Hardness
Tester, DUH-202, Shimadzu Corp., Japan. A Vickers diamond
indenter was pressed onto the flat surface of each sample, until
a complete load-indentation depth-curve was achieved. From
this, the Martens hardness HM could be calculated as
HM ¼ F=ð26:43h2max Þ,
ð1Þ
where F is the test load in (N), and hmax is the indentation depth
under applied test load in (mm).
HM is therefore a simplified hardness value (ignoring the
viscoelastic recovery of the polymer) compared to the classical
indentation hardness, in which the maximum load is related to an
area function. The latter is calculated from the product of a
constant c (dependent on the indenter geometry) and the depth
hc as the intercept of the tangent line drawn from the upper part
of the unloading curve (describing the elastic deformation effect)
and the displacement axis [12].
Scratch Direction
and Speed
Normal Load
Polymer
Sample
Stainless Steel
Indenter
Indenter: Tip Diameter 1 mm
Roughness 60–100 nm
Scratch Length 15 mm
Scratch Speed 10 mm/s
Load Increase from 5 to 90 N
Test Load [mN]
Fig. 2. Photo of the scratch testing device (a), and schematic illustration of scratch test and corresponding testing parameters (b).
1600
1500
1400
1300
1200
1100
1000
900
800
700
600
500
400
300
200
100
0
PBI
PPP
0
1
2
3
4
5
6
PEEK
7
8 9 10 11 12 13 14 15 16 17 18 19 20
Indentation Depth [μm]
Fig. 3. Test load vs. indentation depth during Martens hardness test of PBI, PPP, and PEEK.
K. Friedrich et al. / Tribology International 44 (2011) 1032–1046
2.3. Scratching machine and test conditions
In close relationship to the ASTM and ISO standard for scratch
testing of polymers [13,14], a custom-built scratch machine
(Surface Machine Systems, LLC, Texas, USA) was utilized to perform the scratch tests at ambient condition (Fig. 2). The machine
is capable of recording tangential and normal forces as well as
scratch distance and instantaneous depth experienced by the
stylus. A stainless steel spherical tip with a diameter of 1 mm and
a surface roughness Ra of 60–100 nm was used for all polymers
tested. Scratching was conducted on the smooth surfaces of the
samples, which were pre-cleaned by an air duster. Since the
hardness of the indenter was always much higher than that of
the corresponding polymers, it can be considered as a rigid body
for simplicity. Due to the various dimensions of the as-received
plates, the scratch length was set in all cases at 15 or 20 mm.
A scratch velocity of 10 mm/s was employed for all polymers, and
the normal load in the progressive load scratch tests increased
from 5 to 90 N.
2.4. Scratch damage analysis
Test Load [mN]
After testing, the depth of the scratches h, at the final position
of the indenter was measured (relative to the height level of the
1600
1500
1400
1300
1200
1100
1000
900
800
700
600
500
400
300
200
100
0
1035
normal specimen surface), using in some cases a violet laser
scanning confocal microscope (Keyence VK 9700), but in most
cases a white light profilometer (FTR GmbH) having a lateral
resolution of 1 mm and a vertical resolution of 3 nm. An average of
at least two measured scratch depths is reported in this study.
An ultra-high resolution field emission scanning electron
microscope (FESEM) (Carl Zeiss SMT AG, SUPRATM 40VP,
Germany) was used to investigate further details of the scratch
surface characteristics, operating with 10 kV acceleration voltage,
a working distance of less than 10 mm, and a secondary electron
detector. Therefore, the respective surfaces were sputter-coated
for 70 s with gold.
The scratch width, which was also measured from the laser
profilometer micrographs for the calculation of the scratch hardness Hs [15] is
Hs ¼ FN =As ¼ ð4FN Þ=pw2 ,
ð2Þ
where FN is the normal load, As is the sliding contact area, and
w is the measured residual width at the end of the individual
scratches (on the height level of the normal specimen surface).
The actual width from hill top to hill top, wHH is actually wider
and was also measured from the scratch pictures for comparison.
Also considering the tangential force FT that is continuously
measured during the scratch process, it is possible to define a
PBI
PI 2
0 1
2 3 4 5
6 7 8
PI 10
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Indentation Depth [μm]
Fig. 4. Test load vs. indentation depth during Martens hardness test of PBI, PI 2, and PI 10.
500
450
Martens Hardness [MPa]
400
390
3.4
+ 34
389
+ 5.0
350
300
197
+ 4.2
250
200
150
130
+ 2.3
130
+ 2.2
PI 10
PI 2
100
50
0
PBI
PPP
PEEK
Fig. 5. Martens hardness values of the different polymers tested.
K. Friedrich et al. / Tribology International 44 (2011) 1032–1046
scratch coefficient of friction SCOF as follows [16]:
SCOF ¼ FT =FN :
ð3Þ
A plot of the latter against the sliding distance should approach
an equilibrium value (after a short scratching-in period) as long as
no drastic changes in the scratch mechanisms during the course
of the scratch occur. If, on the other hand, a transition from one
scratch feature (e.g. microcracking) to another (e.g. plowing
associated with material removal) takes place, this should also
be associated with a change in SCOF (as it was shown in Ref. [16]).
An estimation of the specific scratch wear rate wss, can also be
determined from the final parts of the scratches when considering
the loss in volume of the material over an average short distance
of the scratch lave, under an average normal load FNave, that acted
over this distance
wss ¼ DV =ðFNave lave Þ ¼ CSs lave =ðFNave lave Þ,
where CSs refers to the cross-section of the scratch shortly before
its end. The latter can be assumed to have the shape of a segment
of a circle, with a chord length equal to w (residual width) and
Scratch Coefficient of Friction [1]
0.50
PI10_Sample9_Scratch2
90
0.40
80
0.30
70
Scratch Coefficient
of Friction
0.20
0.10
ð4Þ
60
50
Normal Force
0.00
40
-0.10
Force [N]
1036
30
Tangential Force
-0.20
20
-0.30
10
-0.40
0
5
10
Scratch Distance [mm]
0
20
15
Fig. 6. Typical course of the scratch loads and the scratch coefficient of friction vs. scratch distance.
Top View
μm
Topography [μm]
z
Width [mm]
Length Profile x-x
z
μm
Distance [mm]
Distance [mm]
Depth Profile z-z
Topography [μm]
wHH
580 μm
54 μm
Distance [mm]
Fig. 7. White light profilometer results of a typical scratch on a PI 10 surface: (a) top view of the scratch with indication of profiles taken; (b) length profile x–x of the
scratch; and (c) depth profile z–z, with clear evidence of the scratch valley in the middle and hills at both edges of the scratch (wHH ¼ 756 mm).
K. Friedrich et al. / Tribology International 44 (2011) 1032–1046
height equal to the scratch depth h:
2
2
CSs ¼ hð3h þ 4w Þ =ð6wÞ:
ð5Þ
1037
produced. However, no remarkable differences were found for the
two different interval times.
Assuming further that FNave is nearly equal to FN at the end of
the scratch, the specific scratch wear rate reduces to
3. Results and discussion
wss ¼ CSs =FN ,
3.1. Martens hardness of different polymers
ð6Þ
3
ð7Þ
In the present case, some of the residual post-scratch depths
were measured with the laser confocal microscope one day after
scratch testing, but the majority of the scratches was evaluated by
white light profilometry eight months after the scratches were
50
Penetration Depth
Recovery
PI10_Sample9_Scratch2
Scratch Depth [μm]
0
Residual Depth
after Scratch
-50
100
100
90
90
80
80
70
70
60
50
-100
40
Normal Force
Penetration Depth
during Scratch
-150
-200
-250
0
5
10
15
30
60
50
40
30
20
20
10
10
Penetration Depth Recovery [%]
PDR ¼ ðPDRPSDÞ=PD:
Fig. 3 shows the typical hardness curves of three polymers
(PBI; PPP; and PEEK), whereby each curve represents an average
of 10 different measurements. The scatter in each case was very
small. While PBI and PPP have almost the same shape, the curve
for PEEK is much flatter, representing a lower hardness and
modulus of this polymer. However, PEEK is still clearly harder
than the two graphite filled PIs, as shown in Fig. 4. In both cases,
PBI can be considered as a reference curve. An evaluation of the
Martens hardness values from the different curves results in the
following order: PBI¼PPP 4PEEK 4PI 10 ¼PI 2 (Fig. 5). The very
narrow scatter of the measured data is also indicated in this
diagram.
Normal Force [N]
expressed in the units [mm /Nm].
The penetration depth recovery (PDR) in polymers [13] can
also be quantified via the progressive load scratch test. For this,
the penetration depth (PD) during the test must be recorded, and
the residual post-scratch depth (RPSD) sometime after the test
needs to be measured. As a result, the penetration depth recovery
during the scratch process can be obtained by
0
0
20
Scratch Distance [mm]
Fig. 8. Relationship between penetration depth and penetration depth recovery.
PI 2
As Received
Surface
x
x
Scratch Direction
μm
μm
Fig. 9. Laser confocal microscope picture of the scratch profile on material PI 2: (a) top view, (b) 2D picture of the end part of the scratch surrounded by the normal, and
as-received surface; and (c) length profile through the middle of the scratch. (For interpretation of the references to colour in this figure, the reader is referred to the web
version of this article.).
1038
K. Friedrich et al. / Tribology International 44 (2011) 1032–1046
3.2. General scratch behavior
A typical testing curve for polymers under the progressive load
scratch test is shown in Fig. 6. Both the normal and the tangential
(frictional) forces do not increase as perfectly linear, but slightly
progressive. However, there is no point in the curves that can
identify a clear onset of failure. When the normal load has
reached its maximum level and the scratch test comes to a
complete stop, there is always some movement of the indenter
in the backwards direction, which has a minor influence on the
penetration depth at this point (which was neglected in further
considerations). The corresponding scratch coefficient of friction
PI 2
μm
w = 726 μm
h =67 μm
μm
wHH = 793μm
Fig. 10. Depth profile through the scratch on material PI 2.
PI 2
741μm
wHH =741
Scratch Surface
SthSf
Graphite Spalling
Fig. 11. SEM pictures of the scratch on the surface of PI 2: (a) top view with height to height width wHH indicated (the latter is of similar size as measured from the laser
confocal microscope scan); and (b)–(d) magnifications of the surface inside the scratch, with local spots of graphite spalling.
K. Friedrich et al. / Tribology International 44 (2011) 1032–1046
(SCOF), also shown in Fig. 6 exhibit a large magnitude of fluctuation at the beginning of the test, due to the inertial effects of
instantaneously accelerating the scratch head to the designated
scratch speed. After this, the SCOF increases slightly, smoothly
and in a digressive manner to its final level, again influenced at
the end by the back-movement of the indenter.
An evaluation of the scratch produced in this way (here:
material PI 10), using white light profilometry leads to the profiles
shown in Fig. 7. The length profile x–x of the residual scratch
shows a non-linear course, with less material deformation (and
scratch visibility) at the beginning and a progressively increasing
1039
residual scratch depth towards the end of the scratch (Fig. 7b).
At this position the scratch depth is also slightly deeper than
expected from the normal scratch course, due to the back-movement of the scratch head (as mentioned above). Besides, more or
less material was pushed upwards at the edges of the scratch and
in front of the indenter.
The depth profile z–z (Fig. 7c) allows obtaining more precise
information on the remaining (residual) scratch width and depth
at the end of the scratch. In the case of PI 10, the remaining crosssection has a segment of circle shape with a chord length of
580 mm and a height of 54 mm. The CSs appears, however, in a
Normal Surface
PI 10
Height Top
wHH = 735μm
Scratch Surface
Graphite Spalling
Fig. 12. SEM pictures of the scratch on the surface of PI 10: (a) top view with height to height.width wHH indicated (the latter is of similar size as measured from the white
light profilometer scan); and (b)–(d) magnifications of the surface inside the scratch, with local spots of graphite spalling.
Fig. 13. Laser confocal microscope picture of the scratch profile on material PEEK: (a) top view, (b) 2D picture of the end part of the scratch; and (c) depth profile through
the scratch, with an indication of the actual scratch depth h, the width w, and the height to height width wHH.
1040
K. Friedrich et al. / Tribology International 44 (2011) 1032–1046
different shape in the profile shown, due to the different scales of
the topography vs. distance axis.
Fig. 8 shows a comparison of the measured penetration depth
(PD) during scratching with the residual depth (RPSD) eight
months after scratch. Both curves possess a slightly non-linear
shape, with a progressive increase in scratch depth over the scratch
distance. Using Eq. (7), the penetration depth recovery (PDR) is of
the order 80% for the material used in this particular test (PI 10).
3.3. Scratch details of materials tested
3.3.1. Polyimide
The two variants of polyimide (PI 2 and PI 10) differ only in
the amount of graphite incorporated into the polyimide matrix.
This results in smaller differences in the mechanical properties,
so the appearance of the scratches on both variants is also rather
similar. Fig. 9 shows the end region of the scratch at which the
90 N load level was reached. The different colors shown in Fig. 9a
refer to different height levels, with the blue area (inside of the
scratch) being below the normal surface level (orange), and
the red rim around the scratch lying above the level of the
specimen’s surface. Fig. 9b shows a 2D view of the scratch,
following the original scratch direction, so that the deeper valley
and the pushed-up hills at the rim of the scratch are more
visible, relative to the as-received surface of the sample. The
length profile x–x through the scratch and its edges (Fig. 9c)
indicates a slightly rougher appearance of the scratch surface
shortly before the indenter came to a final stop that is associated
with a little backwards movement. Here the profile indicates a
slightly deeper scratch depth than shortly before this point was
reached. It is followed by a steep reduction in scratch depth up
to the heightened rim region, before the normal plateau of the
plate became evident.
A corresponding 2D plot of the scratch with an indicated depth
profile z–z through the deepest point of the scratch is shown in
Fig. 10a. The measured values here (depth h¼67 mm; and width
w¼726 mm) were used for further calculations (Fig. 10b). The
segment of circle shape of the CSs is much more visible in the
present case, as compared to Fig. 7c.
SEM micrographs of the scratches on the material PI 2 (Fig. 11)
and PI 10 (Fig. 12) illustrate that the surfaces inside of the
scratches are relatively smooth. Only higher magnifications of
PEEK
Normal Surface
wHH = 721 μm
Height Top
Scratch Surface
Fig. 14. SEM pictures of the scratch on the surface of PEEK: (a) top view with height to height width wHH indicated (the latter is of similar size as measured from the laser
confocal microscope scan); and (b) magnification of the heightened edge of the scratch, in relation to the scratch and normal surface.
PPP
Scratch Direction
Extrusion Direction
Sharp Features
μm
w = 410 μm
h = 11 μm
μm
wHH = 466μm
Fig. 15. Laser confocal microscope picture of the scratch profile on material PPP: (a) top view; (b) 2D picture of the end part of the scratch; and (c) depth profile through
the scratch, indicating the actual scratch depth h, the width w, and the height to height width wHH.
K. Friedrich et al. / Tribology International 44 (2011) 1032–1046
these surfaces show local sites of material flaking and these
features are more frequently observed for PI 10 than for PI 2
(cf. Figs. 11c and 12c). Therefore, it seems to be reasonable to
assume that these spalling effects are associated with the graphite
flakes filled into the PI matrix for better friction behavior, i.e.,
lower SCOF. When comparing the width of the scratch end
regions (from hill top to hill top) shown in Figs. 11a and 12a, it
becomes obvious that the PI 2 possesses a slightly lower scratch
resistance than PI 10.
1041
3.3.2. Polyetheretherketone
The corresponding scratch pattern on the PEEK surface is
shown in Fig. 13. Both, the width and the depth of the scratch
are clearly smaller than that in case of the PI-samples. Otherwise,
the features look very similar.
This is also true when the SEM of the PEEK scratches is
observed (Fig. 14). No spectacular surface damage features are
visible within the scratch region. Only some pushed-up material
hills clearly indicate the scratch contour.
Fig. 16. SEM pictures of the scratch on the surface of PPP: (a) top view with height to height width.wHH indicated (the latter is of similar size as measured from the white
light profilometer scan); (b) sharp crack features on both side of the scratch; and (c) and (d) magnifications of the periodic cracks, starting from the inside of the scratch
and pointing opposite to the scratch direction.
Fig. 17. Laser confocal microscope picture of the scratch profile on material PBI: (a) top view; (b) 2D picture of the end part of the scratch; and (c) depth profile through the
scratch, indicating the actual scratch depth h, the width w, and the height to height width wHH.
1042
K. Friedrich et al. / Tribology International 44 (2011) 1032–1046
3.3.3. Polyparaphenylene
When it comes to the PPP-samples, the contours of the
scratches are even flatter than that of PEEK, indicating a higher
scratch resistance of this material. Fig. 15 shows that only a
small amount of material is pushed upwards at the edges of the
scratch and in front of it. Some sharp features are visible in
these regions, which may be a result of periodic, and parabolic
cracks pointing opposite the scratch direction (as it has been
observed in previous studies on epoxy and polycarbonate [1]).
But this conjecture can only be validated after higher resolution
SEM analysis.
In fact, the SEM micrographs give a clear evidence for the
existence of such cracks on both sides of the scratch (Fig. 16).
It seems that each pair of the cracks on both sides belongs to one
parabolic crack, of which the middle part was smeared over by a
thin layer of material as the stylus moved passed it in the scratch
direction.
observed, which might have been the reason for the roughened
features seen with the laser confocal microscope.
3.4. Comparison between different materials
3.4.1. Scratch depth, wear rate and SCOF
When considering the residual scratch depth and the measured scratch coefficient of friction, both seem to be related to
some degree. However, this relationship is not linear, but seems
to be a matter of the material’s characteristic. PBI and PPP, both
possessing high modulus, high strength, but a rather brittle
behavior exhibit low scratch depth and low SCOF (Fig. 19). On
the other hand, the more ductile polymers (here PEEK, PI 10, and
PI 2) possess a higher SCOF that is associated with a higher level
90
Scratch Depth h [μm]
80
3.3.4. Polybenzimidazole
It is even more difficult to recognize the scratches on the PBI
samples (Fig. 17). Within the roughness environment on the
surface of the machined samples, the scratch contour is only
visible with the laser confocal microscope due to some pushed-up
material chunks at the edges. Surprisingly, the scratches are
clearly detectable when the samples are observed with the
naked eye.
Also by SEM, the scratch contour on the PBI surface is hardly
visible (dotted line in Fig. 18a). Outside this border line, the
specimen surface is characterized by a granular topography,
which seems to be due to the sintering process of the fine PBI
powder during the compression molding of the plates. Within the
scratch, the surface structure looks a bit smoother, obviously a
result of the ironing effect by the moving stylus. But no deep
scratch effects are otherwise visible. Only at the edges some small
cracking phenomena (pointing in the scratch direction) can be
70
60
PI 10
PI 2
50
PEEK
40
30
20
PPP
10
PBI
0
0
0.10
0.15 0.20 0.25 0.30 0.35 0.40
Scratch Coeff. of Friction SCOF [1]
0.45
Fig. 19. Relationship between scratch depth and scratch coefficient of friction for
the different materials tested.
Fig. 18. SEM pictures of the scratch on the surface of PBI: (a) top view with height to height width wHH indicated (the latter is of similar size as measured from the white
light profilometer scan); and (b)–(d) different magnifications of the boundary between scratch and surrounding, normal material surface, with some indication of short
cracks, starting from the edge into the scratch interior.
K. Friedrich et al. / Tribology International 44 (2011) 1032–1046
of scratch depth. Between the two PI’s, the one with the higher
content of graphite (PI 10) shows a slightly lower SCOF. This
confirms an earlier statement by others, suggesting that a lower
90
80
PI 2
Scratch Depth h [μm]
70
PI 10
60
50
PEEK
40
30
20
coefficient of friction is desirable for polymers to be more scratch
resistant [15].
The depth of a scratch is also associated with a final removal of
material from the scratched surface. The latter can be expressed
by the specific scratch wear rate. Plotting the residual scratch
depth against this wear rate yields an almost linear relationship
between the two quantities (Fig. 20). This is not surprising, since
the scratch depth was also used for the calculation of the scratch
wear rate (according to Eqs. (4)–(6)). In the future we will
compare this scratch wear rate with wear rates obtained from
different tests, such as sliding against smooth steel or abrasive
paper in order to see if any good correlations between these tests
exist. At least from a previously published study by Sinha et al.
[17] it seems that the projected wear debris area after scratching
can be correlated with the specific wear rate found in pin-on-disc
experiments with different polymers.
PPP
10
PBI
0
0
100
150
200
300
250
350
50
„
Spec. “Scratch Wear Rate Wss [10-3mm3/ Nm]
400
Fig. 20. Relationship between scratch depth and specific scratch wear rate for the
different materials tested.
100
Penetration Depth Recovery [%]
1043
90
80
70
3.4.2. Penetration depth recovery
Polymers are viscoelastic and viscoplastic in nature. A portion
of the material at the rear edge of the scratcher recovers almost
instantaneously after the scratcher moves forward [18]. It is
obvious that the more elastic recovery, the better it becomes to
minimize the scratch visibility. The recovery of polymers can be
estimated by the penetration depth recovery upon removal of the
applied scratch stress [19]. A typical testing curve for polymers
under the progressive load scratch test is shown in Fig. 8, along
with the calculated penetration depth recovery. In general, a
relatively high viscoelastic recovery (70%–98%) was found for all
the polymers investigated (Fig. 21). The material parameters that
influence the level of recovery appear to be complex. They are
being subjected to further investigation.
60
50
3.5. Correlations with other material properties
40
30
20
10
0
PBI
PPP
PEEK
PI 10
PI 2
Fig. 21. Penetration depth recovery for the different materials tested.
σB
h
E
[μm] [GPa] [MPa]
90
225
9
80
8
200
70
7
175
An attempt to correlate the scratch results with the mechanical properties of the materials (as listed in Table 1) did not show
any simple relationship (Fig. 22). This was also found out by Sue
et al. [15] who finally concluded that different sets of material
parameters and other relevant factors may be needed to properly
correlate with different scratch damages. This statement was
based on experiments with polymers of different specific
εB
anI
[J/m] [%]
90
90
σB
anI
E
h
80
80
70
70
60
60
60
6
150
50
5
125
50
50
40
4
100
40
40
30
3
75
30
30
20
2
50
20
20
10
1
25
10
10
0
0
0
0
0
εB
PBI
PPP
PEEK
PI 10
PI 2
Fig. 22. Comparison between the residual scratch depth (h) and different mechanical properties (E¼ elastic modulus; sB ¼ strength; anI ¼notched Izod impact energy; and
eB ¼strain at break) of the different materials tested.
1044
K. Friedrich et al. / Tribology International 44 (2011) 1032–1046
90
Table 2
Calculation of the apparent strain e within the final stage of scratching, using a
modified Tabor relationship [23] of the form e ¼0.2a/(R2 a2)1/2, where a¼ w/2
(see Eq. (2)), and R¼ radius of indenter (0.5 mm). Average values were used for the
calculation.
Scratch Depth h [μm]
80
70
60
PI 2
PI 10
50
PEEK
40
PPP
PBI
0
0
2
3
4
5
6
7
Elastic Modulus E [GPa]
8
PBI performance prod.
Celazole U60 SD
Solvay adv. polymers
PrimoSpire PR 120
Victrex PEEK 450 G
DuPont (Vespel)
ST 2002 (2% Gr)
DuPont (Vespel)
ST 2010 (10% Gr)
90
80
70
60
PI 2
PI 10
50
PEEK
40
30
20
PBI
10
PPP
0
0
10
40
50
60
20
30
Notched Izod Impact Energy [J/m]
70
80
90
80
70
60
PI 2
PI 10
50
40
PEEK
30
20
PPP
PBI
0
0
5
10
15
20
25
Strain at Break [%]
30
35
PI 10
h (mm)
w (mm)
e (%)
5.5
242
4.5
10.6
375
8.0
48.3
57.0
538
650
12.7
17.1
52.0
573
14.0
9
Fig. 23. Relationship between scratch depth and elastic modulus for the different
materials tested.
Scratch Depth h [μm]
PBI
PEEK
PI 2
20
10
Scratch Depth h [μm]
Manufacturer company code
PPP
30
10
Short name
40
Fig. 24. Relationship between scratch depth and notch Izod impact energy (a) and
strain at break (b).
performance: (i) ductile and strong; (ii) ductile and weak;
(iii) brittle and weak; and (iv) brittle and strong.
In spite of this fact, a first indication for the scratch resistance
of a polymeric material is the elastic modulus. Sue et al. [15]
showed that for a variety of polymers, a higher modulus was also
associated with reduction in the residual scratch depth. This is
also evident in the results of this study, as shown in Fig. 23.
A similar tendency became visible when plotting the scratch
depth vs. the tensile strength of these materials.
On the other hand, a comparison of the scratch depth with the
two properties representing the notch sensitivity (anl) and the
ductility (eB) tested in this study yielded trends, which hardly
allow any serious conclusions to the materials’ scratch resistance
(Fig. 24a and b). A better correlation can be found, on the other
hand, when comparing the scratch depth with the apparent
strains during scratching. The calculation of which is given in
Table 2 [20].
A plot of the scratch depth and the SCOF against the scratch
hardness, calculated according to Eq. (1) leads to a similar shape
of the curve as it was seen for h vs. E (Fig. 25). Also, this is in good
agreement with previous findings made by Sue et al. [15].
It must be mentioned that the scratch hardness values are a
function of the geometry and material type of the scratch stylus,
but is also dependent on the scratch velocity and testing temperature. Briscoe et al. [21,22] measured values in the range
between 70–800 MPa for various polymers, including polytetrafluoroethylene (PTFE), polyethylene (UHMW-PE), polymethylmethacrylate (PMMA), PEEK and polycarbonate (PC). This means
that the values found here are in a reasonable level, especially
when considering that here we are dealing with the hardest
polymers available so far (i.e., PPP; and PBI).
Finally, it is noted that in the present case a relatively linear
relationship is found when plotting the scratch depth against the
Martens hardness (Fig. 26). Since HM is an easily measurable
value, this parameter can be used to estimate the polymeric
materials’ scratch resistance in terms of residual scratch depth.
However, other works [23] have shown that HM is only useful for
simple surfaces that exhibit simple damage features, such as
ironing. If the surface exhibits skin-core characteristics or possesses a complex morphology, HM would no longer be a good
parameter to correlate with scratch deformation and damage.
This also means if events like cracking and ductile drawing are
present, HM may not be a good parameter to consider (although
in the present case, in spite of crack features in PPP or ductile
plowing events in case of PI and PEEK), the relationship did not
look bad. Nevertheless, significant work is still needed to establish
the fundamental structure-property relationships for the scratch
behavior of high performance polymers.
4. Conclusion
From this study it can be concluded that the scratch behavior
of various high performance polymers follows, in principle, the
K. Friedrich et al. / Tribology International 44 (2011) 1032–1046
1045
90
0.45
PI 2
0.40
Scratch Depth h [μm]
70
0.35
PI 10
60
0.30
PEEK
50
0.25
40
0.20
30
0.15
PPP
20
0.10
PBI
10
0
0
250
500
750
1000
1250
1500
Scratch Hardness HS [MPa]
1750
Scratch Coeff. of Friction SCOF [1]
80
0.05
0
2000
Fig. 25. Plot of scratch depth and scratch coefficient of friction against scratch hardness for the different materials tested.
However, this relationship
is more reliable if all materials
involve the same damage
features and do not have any
effects due to additives
90
80
Scratch Depth h [μm]
70
PI 2
60
PI 10
PEEK
50
40
30
20
PBI
10
PPP
0
0
50
100
150
200
250
Martens Hardness HM [MPa]
300
350
400
Fig. 26. Relationship between scratch depth and Martens hardness for the different materials tested.
same scheme as it has been observed for other commodity or
engineering plastics. However, since the polymers investigated
here possess a relatively high elastic modulus, the scratch
recovery was higher, and in some cases the scratches were more
difficult to detect than in previous studies, where lower performance polymers were tested.
A simple correlation of the scratch resistance with other
material properties could not be found here, i.e., it is a combination of different properties and other factors, which determines how deep the scratches on a particular polymeric
material can be. Although one can state, in general, that a
higher modulus or a higher Martens hardness of the polymer
are favorable for a lower scratch sensitivity. The relative role of
modulus or hardness vs. ductility or impact toughness still
remains unclear. The most scratch resistant material in this
study was polybenzimidazole.
Acknowledgments
One of us, K. Friedrich, is grateful to the German Research
Foundation (DFG FR 675/51-1) for initiating the collaboration
with the colleagues at Texas A&M University, USA. The authors
would also like to acknowledge the Distinguished Scientist
Fellowship Program at the King Saud University, Saudi Arabia,
for the financial support of this project. Special thanks are also
given to Polymer Scratch Consortium at Texas A&M University for
the access and usage of its research facility.
References
[1] Jiang H, Browning R, Sue H-J. Understanding of scratch induced damage
mechanisms in polymers. Polymer 2009;50:4056–65.
1046
K. Friedrich et al. / Tribology International 44 (2011) 1032–1046
[2] Zum Gahr K-H. Microstructure and wear of materials. Amsterdam: Elsevier;
1987.
[3] Wong M, Lim GT, Moyse A, Reddy JN, Sue H-J. A new test methodology for
evaluating scratch resistance of polymers. Wear 2004;256:1214–27.
[4] Wong M, Moyse A, Lee F, Sue H-J. Study of surface damage of polypropylene
under progressive loading. J Mater Sci 2004;39:3293–308.
[5] Xiang C, Sue H-J, Chu J, Masuda K. Roles of additives in scratch resistance of
high crystallinity polypropylene copolymers. Polym Eng Sci 2001;41:23–31.
[6] Chu J, Xiang C, Sue H-J, Hollis RD. Scratch resistance of mineral-filled
polypropylene materials. Polym Eng Sci 2000;40:944–55.
[7] /http://www.celazolepbi.com/files/_u60/CELAZOLE-U-60-PBI-Perf_002.pdfS.
[8] /http://www.solvayadvancedpolymers.com/static/wma/pdf/7/9/8/2/Primos
pire_PR120.pdfS.
[9] /http://www.victrex.com/docs/datasheets-docs/450G.pdfS.
[10] /http://www2.dupont.com/Vespel/en_US/assets/downloads/vespel_s/
H74544.pdfS.
[11] Grellmann W, Seidler S, editors. Munich, Germany: Hanser; 2007.
[12] Briscoe BJ, Fiori L, Pelillo E. Nano-indentation of polymeric surfaces. J Phys D:
Appl Phys 1998;31:2395–405.
[13] ASTM International, ASTM D7027-05; 2005.
[14] International Organization for Standardization, ISO 19252:2008; 2008.
[15] Xiang C, Sue H-J, Chu J, Coleman B. Scratch behavior and material property
relationship in polymers. J Polym Sci B: Polym Phys 2001;39:47–59.
[16] Browning, R, Sue, H-J: Minkwitz, R, Charoensirisomboon, P. Effect of acrylonitrile content on the scratch behavior of styrene–acrylonitrile random
copolymers. Polym Eng Sci, in press. doi:10.1002/pen.22003.
[17] Sinha SK, Chong WLM, Lim S-C. Scratching of polymers—modeling abrasive
wear. Wear 2007;262:1038–47.
[18] Briscoe BJ, Evans PD, Biswas SK, Sinha SK. The hardnesses of poly(methylmethacrylate). Tribol Int 1996;29:93–104.
[19] Jardret V, Zahouani H, Loubet JL, Mathia TG. Understanding and quantification of elastic and plastic deformation during a scratch test. Wear
1998;218:8–14.
[20] Pelletier H, Le Houerou V, Gauthier C, Schirrer R. Scratch experiments and
finite element simulation: friction and nonlinearity effects. In: Sinha SK,
Briscoe BJ, editors. Polymer Tribology. London, UK: Imperial College Press;
2009. p. 108–40. Chapter 4.
[21] Stuart BH, Briscoe BJ. Scratch hardness studies of poly (ether ether ketone).
Polymer 1996;37:3819–24.
[22] Briscoe BJ, Pelillo E, Sinha SK. Scratch hardness and deformation maps for
polycarbonate and polyethylene. Poly Eng Sci 1996;36:2996–3005.
[23] Browning R, Lim G, Moyse A, Sun L, Sue H-J. Effects of slip agent and talc
surface treatment on the scratch behavior of TPOs. Poly Eng Sci 2006;46:
601–608.
Fly UP