Tribological behavior of carbothermoplastic compositesJoël PorretMaster Thesis carried out at Greene, Tweed & Co. (Switzerland) SA in collaboratSurpervisors: Prof. Véronique Michaud (LTC) and Dr. Nicolas Weibel (Greene, TMarch 2013
INTRODUCTIONIn the petrochemical and power industries, a wide range of bearings and wear rings are ubetween high-velocity rotating parts (up to 3600 rpm) and stationary parts. Different materlatter have shown great potential when sliding against metals, because they have excereduce the risk of catastrophic failure of the pump and they allow to diminish the bereduce the risk of catastrophic failure of the pump and they allow to diminish the becomposites, carbon fiber reinforced polyether ether ketone (CF/PEEK) composites and cacan withstand high operating temperatures (up to 260°C) and they exhibit excellent chepurpose of this project was to assess the influence of the matrix nature, the fiber orientCF/PEEK and CF/PFA composites. To do so, Greene, Tweed designed and built a jcounterfaces (journal). In a first phase, dry sliding tests were run, because periods of nbearing durability in absence of any lubricant, high wear resistance and thermal conductivlubricated tests were carried out. First, the goal was to observe the effect of the wetthydrodynamic regimes. Then, the wear resistance of some bearings was evaluated in the
MATERIALS A
Bearings features12 different bearings were produced bydifferent raw materials or by changing thorientation. Their properties are presenthe Table below:
Raw material Bearing names Processingtechnique
CF/PFA calendared paper
CF/PFA paper Comp Compression
CF/PFA paper T3 Techna3™
CF/PFA CF/PFA multiax T3 Techna3™
Bearing manufacturingTwo different molding processes were selected to produce semi-finished parts. Annealing and machining was always required after molding.
Techna3™ process (property ofGreene, Tweed), which allows toproduce tubes (without through-thickness fibers). The material atthe interior of the tube (the partwhich is tested) was often differentfrom the external material, whose
l t i th multiaxial mat CF/PFA multiax T3 Techna3
CF/PFA 3D braid
CF/ PFA Braid 30% Comp Compression
CF/PFA Braid 47% Comp Compression
UD CF/PEEK
UD 0° Techna3™UD 90° Techna3™
UD ±45° Techna3™UD ± 87° Fiber placement
CF/PEEK calendared paper CF/PEEK paper T3 Techna3™
AF (aramid)/PEEK calendared paper AF/PEEK paper T3 Techna3™
CF/PEEK +2% /
role was to increase theresistance and decrease thethermal expansion.
Compression molding. Bothhollow and solid cylinders weremolded depending on the rawmaterial selected. Compressionmolded parts have all the fiberspreferentially oriented in the planeperpendicular to the cylinder axis(most of the fibers have athrough thickness component) PTFE+2% carbon
blackCF/PEEK+fillers Compression
Friction (dry sliding) Bearing wear (dry sliding)
RES
PEEK-based bearings:
The PEEK-based bearings (andcorresponding counterfaces) sufferedsevere wear. Abrasive wear wasdominant mechanism because no tr
through-thickness component).
0.10
0.15
0.20
0.25
0.30
ffici
ent o
f fric
tion
[-]
dominant mechanism because no trfilm was formed on the counterface.
For the UD composites, UD 0° is thewear resistant, followed by UD 45finally UD 90° is the least wear resist
Wear is lower when using aramid(AF) instead of CF because these fibenot abrade steel. However, the finduced by the AF is so high thamaterial is not interesting for beapplications where periods of dry-rube encountered.
Incorporation of fillers (PTFE and gra
PEEK-based bearings: For these bearings, a large amount of debris was
produced, which influenced the friction. However, thereis still a clear effect of the fiber orientation.
For the UD composites, UD 0° shows the lowestcoefficient of friction, followed by UD 45° and UD 90°.
The polymer melts above pv= 0.5 MPa·m/s.
0.00
0.05
pv=0.2 MPa·m/s (p= 0.2 MPa, v=1 m/s)
pv=0.3 MPa·m/s (p= 0.2 MPa, v=1.5 m/s)
pv=0.5 MPa·m/s (p= 0.33 MPa, v=1.5 m/s)
Coef
UD 0° UD ±45° UD 90° CF/PEEK paper T3
0.25
0.30
on [-
]
Thermal conductivity
A linear relationship exists betweenμpv and the steady state temperatureat a certain distance from the bearingsurface. The higher the slope, thehigher the thermal conductivity is. Asexpected, higher thermal conductivities
Incorporation of fillers (PTFE and graallows to promote the formationtransfer film and therefore to drasdecrease wear.
PFA-based bearings: Friction is lower for the bearings having all the fibers in
the plane of sliding (T3) than for those having thefibers out of the plane of sliding (Comp).
0.00
0.05
0.10
0.15
0.20
0.25
pv=0.5 MPa·m/s (p= 0.33 MPa, v=1.5
m/s)
pv=0.7 MPa·m/s (p= 0.33 MPa, v=2.14
m/s)
pv=0.9 MPa·m/s (p= 0.42 MPa, v=2.14
m/s)
pv=0.9 MPa·m/s (p= 0.18 MPa, v=5.10
m/s)
pv=0.9 MPa·m/s (p= 0.08
MPa, v=10.76 m/s)
Coef
ficie
nt o
f fric
tio
CF/PFA paper T3 CF/PFA paper Comp CF/PFA multiax T3CF/PFA braid 30% Comp CF/PFA braid 47% Comp
are measured when the fibers have athrough-thickness component. At highspeeds, the motor cools the housing.
CONCLUSIONSThe formation of a transfer film on the counterface is the key point when carbon fibers arewear in dry conditions. The PFA-based bearings are very promising thanks to their extremfilm, which prevents abrasive wear to occur. Even though both compression molding ofbearings with excellent tribological properties, compression molded bearings showed the b
p g ( p) Friction increases with increasing fiber volume
fraction. The polymer melts above pv=0.9 MPa·m/s.
on fiber-reinforced
tion with the Laboratory of Polymer and Composite Technology (LTC)Tweed)
sed in the pump systems. These are very critical components, since the represent the linkials are still used in practice, such as metals, ceramics and thermoplastic composites. The
ellent mechanical properties, they are able to run dry for extended periods of time, theyearing clearance and therefore to increase the pump efficiency Among thermoplasticearing clearance and therefore to increase the pump efficiency. Among thermoplastic
arbon fiber reinforced perfluoroalkoxy (CF/PFA) composites are very interesting, since theyemical resistance. However, their tribological behavior is not well-known. Therefore, thetation and the fiber volume fraction on the tribological behavior of bearings made out ofournal bearing test rig which allows to test dry and lubricated bearings against steel
non-lubricated contact quite often occur in these pump systems. In order to improve thevity and low friction and thermal expansion are required. In a second phase, some water-ability (PEEK versus PFA-based composites) on the transition between the mixed andmixed lubrication regime.
ND METHODS
y usinghe fibernted in
Vf (%)
Xf (%) Fiber orientation
28 25
Short fibers randomly oriented in the planes
Short fibers randomly oriented in the sliding plane
34 30 Long fibers randomly oriented in
Testing110 dry tests were carried out for a duration of 3 hours andsometimes 6 hours. At least 3 values of pv (the pressure times thevelocity) were applied on each type of bearing. Each testconfiguration was repeated at least 3 times.
34 30 the sliding plane
34 30 Long fibers preferentially oriented in the planes
(due to compression)52 47
61 68
0° in the slinging plane90° (hoop) in the sliding plane
±45° in the sliding plane±87° in the sliding plane
20 25Short fibers randomly oriented in
the sliding plane21 25
Short fibers randomly oriented in
The shaft (journal) was made out of 304 stainless steel and a newpart was used for each test. Regarding the dry-running tests, theshaft diameter measured 68.5 mm and the bearing inner diameterwas set to 70 mm For the lubricated tests the shaft diameters
Housing 3D model of the test setup
Steel counterface (journal)Journal bearing test rig
25 32 Short fibers randomly oriented in the planes
ULTS
theird froms theansfer
PFA-based bearings:
Wear is up to 2 orders of magnitude lowerwhen a transfer film is formed on thecounterface.
The PFA based bearings were always able
was set to 70 mm. For the lubricated tests, the shaft diameterswas of 70 mm and the bearing inner diameter measured 70.2 mm.
1.5E-04
2.0E-04
2.5E-04
3.0E-04
wea
r rat
e [m
m3 /
Nm
]
UD 0°
UD 45°
UD 90°
No transfer film on the counterface
ansfer
e most5° andtant.
fibersers dofrictionat thisearing
un can
aphite)
The PFA-based bearings were always ableto form a transfer film.
Wear is higher for the bearings having allthe fibers in the sliding plane (shearingbetween the layers) than for those havingthe fibers out of the plane of sliding.
Wear slightly decreases with increasingfiber volume fraction (for the braids).
Improving the adhesion fiber/matrixincreases the wear resistance.
At high velocities (3000 rpm), the bearingshaving all the fibers in the sliding plane
4.E-06
5.E-06
6.E-06
7.E-06
8.E-06
9.E-06
ear r
ate
[mm
3 /N
m]
CF/PFA paper T3
CF/PFA paper Comp
CF/PFA multiax T3
/
0.0E+00
5.0E-05
1.0E-04
Spec
ific
w
CF/PEEK paper T3
AF/PEEK paper T3
Transfer film on the counterface
Water-lubricated tests
There is no effect of the wettability(wall slip) on the transition betweenthe mixed and hydrodynamicregimes. This is probably due tothe fact that the bearing is thestationary part. In the mixed
aphite)of a
stically
g s s g psuffer from severe wear.
Wear increases with increasing velocityand decreasing pressure (same pv value).0.E+00
1.E-06
2.E-06
3.E-06
Spec
ific
we
CF/PFA braid 47% Comp
CF/PEEK+fillers (transfer film)
6080
100120140160180200220
Tem
pera
ture
[°C]
CF/PFA paper T3CF/PFA paper CompCF/PFA multiax T3CF/PFA braid 30% CompCF/PFA braid 40% Comp
3000 rpm
1423 rpm
00.020.040.060.08
0.10.120.140.160.18
1.E-06 1.E-05 1.E-04 1.E-03
µ [-]
Transition
regime, no wear was observed forall the bearings tested.
e used as reinforcement. Without fillers, the PEEK-based bearings studied exhibit severemely low friction against steel and their ability to form a thin and strongly bonded transferCF/PFA composites and co-molding of CF/PFA on CF/PEEK composites allow to obtain
best comprise between friction, wear, thermal conductivity and thermal expansion.
60
0.08 0.13 0.18 0.23
μ·pv [MPa·m/s]
ηN/P [-]CF/PFA paper Comp. Ra = 0.4 µmCF/PFA paper Comp. Ra = 2 µmCF/PFA multiax T3. Ra = 0.7 µmCF/PEEK+fillers. Ra = 1 µm