Sem Aluminium

Published on November 2016 | Categories: Documents | Downloads: 39 | Comments: 0 | Views: 202
of 23
Download PDF   Embed   Report

sem images of aluminum alloys

Comments

Content


Wear, 69 (1981) 1 - 23
@ Elsevier Sequoia S.A., Lausanne - Printed in The Netherlands
1
TOPOGRAPHICAL FEATURES OBSERVED IN A SCANNING
ELECTRON MICROSCOPY STUDY OF ALUMINIUM ALLOY
SURFACES IN SLIDING WEAR
J. CLARKE and A. D. SARKAR
J ohn Dalton Faculty of Technology, Manchester Polytechnic, Chester Street, Manchester
Ml 5GD (Gt. Britain)
(Received March 28,198O)
Summary
Extensive examination of sliding wear surfaces was undertaken on
various Al-Si alloys using the scanning electron microscope in an attempt to
elucidate the prevailing mechanism or mechanisms of wear. Recurring surface
features were noted which gave a positive indication of diverse mechanisms
operating across the range of material composition and contact stresses
involved.
It is postulated that mutual transfer across the wear interface is an
extremely significant feature of all regimes of sliding wear, becoming more
prevalent in materials whose yield stresses are low or where normal contact
stresses are high. Asperity interaction, gross layering and ductile fracture
constitute the predominant mode for conditions bordering on seizure,
whereas a shear transfer mechanism operates in regimes between those of
mild and severe wear. Delamination producing wear craters is a predominant
feature of mild wear regimes in eutectic and hypereutectic alloys, but
primary and secondary modes dependent on subsurface and surface failure
respectively can be recognized.
The concept of mutual transfer suggests that pin surface topography is
a function of material back transfer such that a new delamination mode can
be recognized where failure results from surface-initiated cracks extending
down to interlayer zones.
Work on asperity macromodels using wax as the simulative medium
provides strong evidence for an almost universal transfer mechanism.
1. Introduction
Interest in observing worn surfaces has increased, and the scanning
electron microscope has made an invaluable contribution towards a basic
understanding of wear processes in materials. The concept that components
wear by a fatigue mode or by simple abrasion is giving way to statements
2
that, in industrial machinery or laboratory test rigs, wear debris are generated
by a number of apparently independent mechanisms although variables such
as speed, load and cleanliness of the interface are constant. Thus a component
may wear partially as a result of delamination [l] which starts with the
postulate that the subsurface is plastically deformed repeatedly during slid-
ing. This results in the formation of subcutaneous voids which elongate in
the direction of shear deformation, leading to the formation of a crack
which eventually extends to the surface. A wear sheet is then lifted from the
matrix and scanning electron microscopy (SEM) reveals that the sliding face
of the debris is smooth but the underside shows shear dimples, indicating a
ductile fracture. Rather than the original matrix material undergoing wear
in a direct involvement, evidence seems to be gathering that the bulk of the
wear product is generated from material transferred onto opposing surfaces.
X-ray microprobe analysis has revealed [2] that couples such as Cu-Fe,
Cu-Zn and Fe-Ni show mutual transfer of metals, i.e. copper on iron and
vice versa for a Cu-Fe couple. Material is transferred by adhesive forces [ 31
and the events take place on an asperity scale [4]. Transferred material is
built up in stages and some detaches as a wear particle by a fatigue mode [ 51
or by delamination as proposed by the delamination theory. SEM studies of
debris confirm a fatigue type of failure [4] or ductile fracture. It is unlikely
[6] that wear debris will necessarily be generated as one asperity adheres to
another occupying a position on the counterface, and it now seems certain
that the predominant event in sliding is mutual transfer and build-up of
layers of asperities. This can be designated as back transfer but the problem
of obtaining positive proof remains. In the following, selected results from
SEM studies of a few hundred wear scars, mainly of binary Al-Si alloys
sliding on hard steel, are discussed. An attempt was made to show that
mutual back transfer across the interface takes place and the bulk of the
wear debris results from these deposits. Although still speculative, some
indirect evidence is provided on the origins of cracks. This is important
because some theories argue forcefully that cracks originate below the surface.
However, the suggestion of surface cracks propagating towards the matrix
cannot be dismissed. An argument [ 71 is developed that the effect of sliding
under load is to form a thin white layer of plastically strained metal. Cracks
form in this layer originating at the surface where the tensile stress is greatest.
When two such cracks intersect below the surface, a wear particle forms by
shear in the region of maximum hertzian shear stress.
2. Results
The results that we report are mainly from SEM studies of aluminium
alloy pin surfaces but observations on the counterface are also summarized.
The counterface in all cases was a hard steel bush which provided a surface
speed of 1.96 m s-l. All sliding wear experiments were carried out in room
atmosphere.
3
2.1, Pin surface
2.1.1. Roof tile laminates
Roof tile laminates are so named because of the stepped appearance
observed when the pin surface is viewed at relatively low magnifications
(Fig. l(a)). Examination of the edges of the steps reveals that these are
made up of a distinct series of layers (Fig. l(b)). The feature is associated
with high interfacial stresses and in its extreme form covers the whole surface
(Fig. l(c)), indicating that seizure is imminent. A milder form of the same
feature is seen in Fig. l(d) where the roof tiles are seen to be in local areas
only. Corresponding features on the counterface are shown in Figs. l(e) and
l(f)*
Explanations pertaining to the mechanism of formation of these lami-
nates might be as follows.
(a) Extreme surface conditions created by high interfacial stresses and
temperatures cause bulk welding of the opposing surfaces, thus necessitating
ductile fracture for sliding to be continuous.
(b) A metal-cutting action may be at work, since Fig. l(c) is reminiscent
of certain magnified machined surfaces involving a built-up edge and shear
front layers.
(c) Asperity interaction concurrent with mutual transfer is responsible
for the directionality and layering observed.
I n the first case, although it must be admitted that adhesion and subse-
quent fracture play a part in the mechanism of formation, it is difficult to
concede that some fracture mode is solely responsible for its occurrence.
A series of cleavage steps might be envisaged, but these would normally be
associated with a brittle material, which this is not, and comparison with
fracture surfaces of this mode does not give the impression that the roof
tiles possess an orientation relationship between the opposing surfaces and
the crystal structure. Ductile fracture modes showing typical shear dimples
also offer similarities of topography but these are usually observed at much
higher magnifications than those for the roof tiles. I n addition, the laminated
structure of the roof tiles defies explanation along these lines. I n the second
case the concept of a metal cutting action is probably more acceptable since
surface features are really on a macroscopic level as will be the roof tile
laminates in their extreme form.
Two difficulties arise with this. One problem is that, if this were taking
place on a large scale sufficient to account for the feature, the wear debris
would inevitably consist of many chips, as formed in a cutting process.
Admittedly there were occasional tiny particles of debris exhibiting the
classic shear front formation, but the prevailing debris form was the typical
mixed lumpy and flat sheet confiirations commensurate with the magnitude
of interfacial stress involved.
The second difficulty is ascertaining what precisely acts as the cutting
tool. On a macroscale, if the pin is strain hardened it is possible that it may
remove the deposited metal by cutting the transferred layer of pin material.
On a microscale it might be considered that hard asperities favourably
(a)
Fii. 1. (a) A&-6.2%& pin surface; load, 2.5 kgf (magn~cation, 140X). (b) Al-6.2%Si pin
surface;load, 2.5 kgf; the direction of counterface motion is into the paper (magnification,
700x). (c) Al-6.2ZSi pin surface; load, 2.6 kgf (magnification, 28x). (d) Al-6,2%Si pin
surface; load, 1.5 kgf (msgnification, 270X ). (e) Counterface surface; Al-6.2%Si pin;
load, 2.6 kgf (magnification, 2700x ).(f) Counterface surface; Al-6.2%Si pin; load, 2.5
kgf (magnification, 360x). The arrows indicate the direction of counterface motion.
5
oriented and of a suitable shape might act in this manner. A further difficulty
again arises in attempting to explain the laminated structure since the layers
(Fig. l(b)) do not have the appearance of the deformation boundaries some-
times observed in metal cutting.
In the third case an explanation might be found in terms of asperity
interaction. It could be argued that the asperities on both surfaces flow
plastically owing to the severe surface stress. If the asperities idealized as
conical in shape were flattened and smeared in the direction of sliding on
their parent surfaces, they would have the direction and form shown in
Fig. 2(a).
Comparing this with the observed direction for the pin and counterface
represented in Fig. 2(b), it is seen that this explanation will not suffice since
the direction of the roof tiles is the wrong way round. Where this feature
appeared, both the pin and the counter-face showed a relatively uniform
distribution of what may be termed hyperbolic profiles of apparently
flattened asperities, as shown schematically in Fig. 2(b). In the majority of
cases the observed shape of the roof tile laminates was actually hyperbolic.
Thus, comparing Figs. 2(a) and 2(b) it is apparent that the idea of asperity
smearing from parent surfaces does not fit the case, unless for some reason
rotation takes place through 180” after initial formation.
This is very unlikely and therefore it is suggested that a mutual transfer
mechanism is operating. Thus, initially, severe plastic flow deposits the
aluminium alloy onto the steel counter-face by means of asperity flattening
and transfer. Some of the deposit transfers back onto the pin surface by the
same mechanism. This results in the directionality of the profiles in reverse
order, as shown in Fig. 2(b).
Evidence for back transfer seems conclusive when Fig. 3 is considered.
Figure 3(a) shows an aluminium pin surface after sliding on a steel counter-
x’ x
Y’ Y
View A WI View B on
munterfnce Dill surface
(a) (b)
Fig. 2. (a) Schematic representation of asperity interaction;(b) orientation and directional-
ity of roof tile laminates at the wear interface.
Fig. 3. (a) Pure aluminium pin on a clean steel track; load, 8 kgf (magnification, 336x).
(b) Pure aluminium pin on its own deposit; load, 8 kgf (magnification, 840x). The arrows
indicate the direction of counterface motion.
face and Fig. 3(b) illustrates a pin surface after sliding on a steel counterface
possessing a track of deposited pin material. Since it is clear that material has
been transferred from pin to counterface, the existence of identical topog-
raphies on pin and counterface suggests that the same mechanism has produced
both. Thus it is a valid assumption that, if the counterface topography is a
result of some transfer mechanism, then that of the pin surface is equally a
function of the same mechanism.
From Figs. l(b), l(e) and l(f) the laminations appear to be very system-
atic. There are about six to ten layers making up the composite roof tiles.
This range was fairly consistent and suggests that transfer as a single event is
followed by successive deposits until the layers are removed as a wear product.
It is possible that some critical number of layers needs to be deposited before
a wear product is produced. A peculiarity is that the laminates are bent back
as they are removed as a result of sliding (Fig. 4). The removal process can
Fig. 4. AL6.2%Si pin surface; load, 2.5 kgf. The arrow indicates the direction of counter-
face motion. (Magnification, 1200x.)
7
evidently be explained by an adhesive shear process as noted in the ductile
shear fracture dimpling on both surfaces of the friction system.
Figure l(b) shows positive divisions between the layers. This would be
so, because the interlayer cold welds will only form a fraction of the apparent
area. The true area of contact may also be diluted because of oxidation or
the presence of non-metallic constituents such as silicon.
A study of surfaces near to seizure is considered to be justified on the
assumption that prevailing wear mechanisms may well be accentuated in
these extreme cases, throwing light on those operating in milder regimes,
the most probable situation in properly operated machinery.
2.1.2. I nclined shear plates
A feature best described as inclined shear plates was only observed in
the combinations of load and composition where plastic flow was evident.
It appeared across the whole composition range but was only rarely seen on
the eutectic or near-eutectic alloy surfaces. It is perhaps significant that the
feature was not normally observed on surfaces showing the roof tile lami-
nates, i.e. where gross plastic flow was evident. It appears to be a kind of
transition feature between the relatively wear-resistant eutectic or near-
eutectic alloys and those which show severe surface flow such as pure
aluminium or alloys with a low silicon content.
Figure 5(a) is an example of its occurrence on the counterface and
shows clearly the nature of the individual plates which incline to the direc-
tion of the tangential force. Identical forms of this feature were observed on
the corresponding pin surfaces. One peculiarity seen in Fig. 5(a) is that the
inclined plates are sited below two smooth deposits on each side of them.
In most cases the feature appeared as shown in Fig. 5(b) and magnified in
Fig. 5(c).
Studies using a conical macroscopic model of wax to observe pin-
counterface deposition modes showed a topography almost identical with
that observed on the real surfaces. Figure. 6 shows wax plates deposited at a
fairly consistent angle and plate distance. Part of the effect is muted by the
fact that wax begins to soften under the photographic light but close observa-
tion during actual sliding showed that the plates were sheared off the wax
pin at the trailing edge with simultaneous deposition onto the counterface.
This evidence and also that its occurrence on the counterface axiomatically
implies a transfer mechanism lead again to the suggestion of back transfer
where the feature is observed on the pin surface.
Further visual evidence exists to support this view. A joint study of
Figs. ‘7(a) and 7(b) showing pm and counterface surfaces respectively suggests
that in both cases a shear transfer mechanism has been responsible for their
formation. It is possible that favourably shaped and oriented asperities on
the respective surfaces have experienced simultaneous shear and deposition
as noted in the macroscopic models. Figure 7(c) is a higher magnification
view of Fig. 7(a) and shows clearly the nature of the deposited plates.
8
Fig. 6. (a) Counterface surface; pure alumin-
/
ium pin; load, 0.75 kgf (magnification,
609x). (b) Pure aluminium pin surface; load,
0.75 kgf (m~nifieation, 234x). (c) Pure
aluminium pin surface; load, 0.75 kgf
(magnification, 650x). The arrows indicate
the direction of counterface motion,
Fig. 6. Wax deposit on steel; load, 1.0 kgf. The arrow indicates the direction of counter-
face motion. (Magnification, 27x.)
(4
(cl
(b)
Fig. 7. (a) Pure aluminium pin surface; load, 0.76 kgf (magnification, 700x). (b) Counter-
face surface; Al-6.2%Si pin; load, 1.5 kgf (magnification, 1000x). (c) Pure aluminium pin
surface; load, 0.75 kgf (magnification, 2800x). The arrows indicate the direction of
counterface motion.
Since real asperities probably vary in shape and orientation, this fact
may explain the peculiarity noted earlier (Fig. 5(a)) where the inclined shear
plates have material built up on either side of them. The suggestion is that
adjacent asperities or asperity groups will yield different deposits since these
will be a function of asperity shape. A similar effect is noticed in Fig. 8(a).
2.1.3. Adhesive delamination
At low loads on a pure aluminium pin or at high loads on certain hyper-
eutectic alloys, shallow surface craters were apparent, giving the appearance
that material had been delaminated from them. Thus the feature was associ-
ated with plastic flow and adhesion effects. Figures 8(a) and 8(b) are examples
(cl
-
(d)
-
Fig. 8. Pure alumiuium pin surface; load, 0.75 kgf. The arrows indicate the direction of
counterface motion. (Magnifications: (a) 275~; (b) 700x;(c) 65X;(d) 250X.)
of the cratering effect, showing diverse topographies for the actual crater
bases. The laminated nature of the crater sides is noted in Fig, 8(b). In view
of the earlier comments on material transfer a mechanism of formation
might be as follows.
Progressive deposition of material onto the pm surface is taking place
simultaneously with material loss, as either loose wear debris or redeposition
onto the counterface. Because of the different shape-orientation features of
the counterface asperities, the pm surface at any instant will consist of
inclined shear plates and smooth deposits. Subsequent deposition may then
layer material over the initial surface either completely or partiaily. The fact
that inclined shear plates did exist at different levels on the surface seems to
support this view.
11
The existence of multilayering is also observed in Fig. 8(b) and it is
probable that the layers are held at isolated points through adhesion, but
the interlayer zones can mainly be regarded as subsurface cracks. It is not
unreasonable to speculate that at some critical stress level the surface fails
at a particular point, resulting in a transverse crack (Figs. 8(c) and 8(d));
then subsequent cycling may cause material to become progressively delami-
nated by adhesive forces peeling the surface off in thin layers. Figure 8(a)
shows a thin peninsula of material, presumably ready for detachment. The
topography of the crater bases would then be simply a function of asperity
shape and orientation which would determine the deposition pattern on the
penultimate or even earlier layers. Figure 9 is a schematic representation of
surface layer removal once transverse crack initiation has taken place. The
depth of the crack may be through one or more layers. Since debris examina-
tion for these cases did reveal very thin sheets, it is probable that two types
of delamination mechanism are at work, one dependent on subsurface cracks
and one where a surface-initiated crack meets with the interlayer zones
between deposited layers.
2.1.4. Granular delamination
Granular delamination occurred exclusively in eutectic and hyper-
,eutectic alloys and, since the incidence of cratering tended to diminish with
increasing load, the feature can be associated with mild wear regimes. Figure
10(a) illustrates the granular nature of the delaminated areas which appeared
in random form across the whole wear surface. A single crater does not
Potential crack
(b)
’ Layer peeled off
surface by adheswe
forces
Fig. 9. Schematic representation of surface layer removal by adhesive delamination: (a)
stage 1; (b) stage 2.
Fig. 10. AI-Pl%Si pin surface; load, 1.0 kgf. The arrows indicate the direction of counter-
face motion. (Magnj~catio~: (a) 368x;(b) 920x.)
indicate the size of a single detached wear plate or sheet since it is obvious
that craze cracking (Fig. 10(b)) occurs around the perimeter, resulting in a
progressive or secondary type of delamination. This secondary delamination
may be a function of surface forces only and may have nothing to do with
the probable subsurface cracking responsible for primary delamination. A
number of observations were made of subsurface cracking, which may be
called primary delamination, typical examples being shown in Fig. 11. In
both cases, but more so in Fig. 11(b), it is interesting to note how the
incipient delamination appears to have been initiated at a transverse crack
which according to Suh [8] would have been of subsurface origin, Instances
of what might be termed secondary delamination were numerous, examples
being seen in Fig. 12. On a purely visual basis it appears that, once the initial
fracture has taken place, subsequent surface deterioration occurs in a series
of delaminations which may well be a result of the action of surface forces
only on the already weakened areas.
Fig. 11. (a) Al-lG%Si pin surface; load, 3.0 kgf (magnification, 240x). (b) Al-2l%Si pin
surface; load, 1.0 kgf (magnification, 200x). The arrows indicate the direction of
counterface motion.
(4
Fig. 12. (a) Al-15.5%Si pin surface; load,
(b)
-
1.5 kgf (magnification, 130x). (b) Al-13%Si
pin surface; load, 1.0 kgf (magnification, 700x). The arrows indicate the direction of
counterface motion.
Subsurface studies did reveal cracks parallel to the wear surface in some
of the alloys and it may well be that these are responsible for the initial
delamination. According to the delamination theory the crack terminates at
the surface and presumably defines the crater perimeter; hence secondary
types of delamination would not normally result as a feature of the initial
subsurface crack. A study of the granular craters does not provide any classic
evidence of fracture as noted by Jahanmir et al. [l] . The granules appear to
be heavily compounded eutectic material and indeed Fig. 13 tends to confirm
this where the eutectic is seen to be compounded and strung out in the direc-
tion of sliding. The laminated effect is of interest, suggesting a subsurface
crack effect. Parallel studies on pure copper showed a granular crater effect
but the granules were not as prevalent as those for the Al-Si samples.
Fig. 13. Al-ll%Si pin surface; load, 1.0 kgf. (Magnification, 1000x.)
14
A preliminary summary of this feature based on the foregoing comments
suggests that two delamination modes may well have to be recognized.
Primary del~ination occurs where initial surface rupture takes place by the
initiation of a subsurface crack, and secondary delamination is experienced
where progressive surface rupture takes place probably as a result of surface
forces only. The surfaces exhibiting this feature appeared to be oxidized and
the abundance of craze cracking (Fig. 14) indicates that a fatigue mode is
probably also operational.
-
Fig. 14. Al-Sl%Si pin surface; load, 1.0 kgf. The arrow indicates the direction of counter-
face motion. (Magnification, 2000x.)
2.1.5. Cracking
A form of failure to be described simply as plough cracking (Fig. 15(a))
was observed mainly with hypereutectic alloys at low contact stresses. The
(a) (b)
Fig. 15. (a) Al-15.5%Si pin surface; load, 1.5 kgf (magnification, 2208x). (b) Al-lG%Si
pin surface; load, 3.0 kgf (magnification, 2400x). The arrows indicate the direction of
counterface motion.
15
surfaces show parallel furrows, creating the impression of a ploughed field;
the cracks are associated with the raised portion of the furrow. Figure 15(b)
shows a typically cracked region but inclined shear plates are also noted
which suggests back transfer from the counterface. The feature gives the
appearance of a brittle shell of material on an unsupported underside which
has cracked and is starting to disintegrate under the repetitive stresses involved
in sliding. The cracks are actually very small and only appear at mag$fica-
tions in excess of 1000X. A more appropriate title might therefore be hairline
plough cracking.
The second type of cracks observed were those appearing transverse to
the sliding direction. They differed from the previous type in that these were
mainly longitudinal to the sliding direction and of a much finer nature. Figure
16(a) is a typical example of transverse cracking and it is interesting to note
(b)
Fig. 16. (a) Al-lB%Si pin surface; load, 0.5 kgf (magnification, 280x). (b) Al-6.2%Si pin
surface; load, 0.5 kgf (magnification, 600x). (c) Al-6.2ZSi pin surface; load, 0.5 kgf
(magnification 2400x). The arrows indicate the direction of counterface motion.
16
how the cracks appear in a kind of series. A similar effect is noted in Fig.
16(b) where the series finally results in loose debris (Fig. 16(c)). A study of
Fig. 16(b) strongly suggests that the cracks have formed in a deposited layer
where surface-initiated cracks terminate at the underlying layer. A number
of examples support this view and the thought is analogous to that of adhe-
sive delamination where loss of material from the surface is a function of
surface cracking in back-transferred layers.
Figure 17(a) shows a transverse crack on a surface associated with
granular delamination. At the top right-hand side of the field, magnified in
Fig. 17(b), is an apparent series of transverse cracks. Closer examination
shows that these closely resemble the inclined shear plate feature. This,
together with the obvious layering noted at the top of the field, confirms the
theme that a thin back-transferred layer is present. In addition, the large
transverse crack may well be the incipient stage of formation for a granular
wear crater which supports the contention already made when we considered
this feature. It is also worthy of note that transverse cracking appeared across
the whole range of alloy composition used in these sliding wear tests.
A related feature under this heading is that of craze cracking particularly
prevalent in the hypereutectic alloys. Figure 14 illustrates typical surface
d~inte~ation by this mode. Figure 18 is of interest since craze cracking has
taken place in a back-transferred layer. Figure 10(b) shows the cracking to
be prevalent in the material around crater perimeters and it appears to have
been initiated at the surface. It is significant that the ll%Na-modified alloy,
which had the lowest wear rate of the alloy series, showed hardly any craze
cracking compared with say the 21%Si alloy with its inferior wear properties.
It may be that the structured substrate (Fig. 13) for the eutectic alloy gave
better surface support than the crushed random composite of primary silicon
and eutectic material in the 21%Si alloy.
The feature was observed across the full range of composition but tended
to be restricted to lower contact stresses.
Fig. 17. Al-13%Si pin surface; load, 1.0 kgf. The arrows indicate the direction of counter-
face motion. (Magnifications: (a) 860X;(b) 4650X.)
17
-
Fig. 18. Pure aluminium pin surface; load, 0.75 kgf. The arrow indicates the direction of
counterface motion. (Magnification, 270X.)
2.1.6. Subsurface caverns
Subsurface caverns are worth noting because typically they appear as
cavities (Fig. 19(a)) which seem to arise because a layer covering them has
been removed by some mechanism. The suggestion that these may be due to
porosity left in the cast pins was discarded because the foundry technique
was specially arranged to avoid porosity due to gas or shrinkage, and samples
scanned under the optical microscope even at a magnification of 500X showed
no porosity. Considering Figs. 19(a) and 19(b) and bearing in mind the
magnifications involved, it is a reasonable suggestion that the cavities appear
as a result of back transfer. Thus, because the contact stresses are low the
layers do not adhere effectively to the substrate. When tangential or other
(4
-
Fig. 19. (a) Al-13%Si pin surface; load, 2.5 kgf (magnification, 1200x). (b) Al-15.5ISi
pin surface; load, 1.5 kgf (magnification, 4680x). The arrows indicate the direction of
counterface motion.
18
(b)
I
(d)
Fig. 20. (a) Al-21%Si pin surface; load, 3.0 kgf (magnification, 5950x). (b) The counter-
face corresponding to (a) (magnification, 696~). (c) Counterface surface; Al-&2%Si pin;
load, 2.6 kgf (rn~i~~tjon, 100X). (d) The pin surface corresponding to (c) (magnifica-
tion, 3000x). The arrows indicate the direction of counterface motion.
19
forces lift part of the layer, the exposed surface disintegrates locally and
reveals a cavity resembling a subterranean cavern.
2.1.7. Shear dimpling
Figure 20(a) shows that the worn surfaces possess small volcanic-type
craters which resemble dimples. They are characteristic of ductile fracture
and in this case are elongated in the direction of sliding. This can happen in
any composition range but is characteristic of heavy loads bordering on
seizure conditions. A surface such as this gives clear evidence that adhesion
and cold welding of junctions have taken place between the pin and counter-
face such that the only way to separate them is by fracture of the junctions.
Because of the dynamic situation existing during the sliding process, it is
axiomatic that such welding and fracture would take place throughout the
operational life of a friction couple, giving rise to stick-slip phenomena and
vibration of the machine. These shear dimples are associated with the roof
tile features and Fig. 20(a) is a view taken directly on top of one of the
steps. The arrow in Fig. 20(b) shows the approximate position of the dimpled
area. It is seen that the dimples are elongated in a direction away from the
step edges, i.e. in the direction of rotation of the counter-face. Although the
roof tile profiles are opposite on the counterface (Fig. 20(c)), the same
elongation relationship exists, as shown in Fig. 20(d). The arrow in Fig. 20(c)
shows that in this case the dimpling is located at the side of the roof tile
deposit.
Figure 21 is a schematic representation showing the relative profile-
elongation relationship. Again, identical relationships on pin and counterface
suggest that mutual transfer has taken place which in view of the evidence
seems to be a more reasonable explanation than that of unidirectional transfer
combined with bulk welding and fracture. It should be noted that the opposite
roof tile profiles are not considered as superimposed during sliding. These
Pin roof tile
Direction of
Dwction of
Counterface
roof tile
Fig. 21. Relationship between roof tile directionality and the elongated fractured dimples.
20
will have formed independently and represent separate transfer events. This
fact tends to reinforce the mutual transfer postulate in that some kind of
bulk cleavage fracture mode would have the same profile directionality even
though the matching surface levels would be different. It is also of interest to
note that void coalescence prior to fracture must be extremely rapid in view
of the surface speeds involved during sliding.
2.2. Counterface observations
Previously published work [9] has shown that the counterface gains in
weight as a result of transfer. This is lost in stages and the counterface itself
wears in certain cases. A thorough examination of the counterface was there-
fore undertaken.
Certain features have already been noted; these are the existence of
inclined shear plates and roof tile laminates in a form identical with those on
the pin surface. Figure 5(a) shows the inclined shear plates with their counter-
parts on the pin surface in Fig. 22(a). The roof tile laminates can be seen in
Fig. 22(b) with their counterparts on the pin surface in Fig. l(c). Some other
features were craze cracking, transverse cracking and, in odd cases where
both silicon contents and contact stresses were high, granular wear craters.
None of these three, however, was prevalent to the same degree as those
observed on the pin surface.
The steel substrate itself wears, as seen previously by weight loss mea-
surements. Microprobe analysis of wear debris showed iron as a definite
constituent well in excess of the original analysis levels. Iron was also noted
in the pin surface. The possibility is that the aluminium wears the steel
surface, which is probable in view of the abrasiveness of the silicon particles
and is confirmed because the hypereutectic alloys gave the highest counter-
face wear rate. Part of the worn steel appears as wear debris and the remainder
becomes embedded in the aluminium pin surface. Iron could also diffuse
into the body of the Al-Si alloy, the process being facilitated by the high
interfacial temperatures of the friction couples. The steel surface itself shows
(a) (b)
Fig. 22. (a)Pure aluminium pin surface; load, 0.75 kgf (magnification, 596x). (b) Counter-
face surface; Al-2l%Si pin; load, 3.0 kgf (magnification, 170x). The arrows indicate the
direction of counterface motion.
Fig. 23. Counterface surface; pure aluminium pin; load, 2.0 kgf. The arrows indicate the
direction of counterface motion. (Magnifications: (a), (b) 1700x; (c) 550x.)
a kind of craze cracking and cratering (Figs. 23(a) and 23(b)) which probably
occurred in the early stages of sliding. In both these cases the aluminium
deposit was removed by dissolving with NaOH. It is probably fair to say that
the counterface wears by some direct effect due to sliding against the Al-Si
alloy pin until a deposit of the Al-Si alloy builds up. Patches of the deposit
may be removed during sliding and the steel surface is then expected to wear
by a direct effect (Fig. 23(c)).
Delamination is illustrated in Fig. 24; Fig. 24(b) almost certainly illus-
trates that delamination has taken place by the action of surface-initiated
forces on the transferred layer.
3. Discussion
It is clear that throughout the foregoing observations one theme has
been dominant, i .e. the idea of mutual transfer where material is transferred
across the interface with some of this deposit transferring back again to the
parent surface. It is thought that some equilibrium transfer rate is established
22
Fig. 24. (a) Counterface surface; Al-13%Si pin; load, 3.0 kgf (magnification, 2670x).
(b) Counterface surface; Al-3%Si pin; load, 3.0 kgf (magnification, 2670x). The arrow
indicates the direction of counterface motion.
with the balance of transferred material being in the pin-counterface direc-
tion. Figure 25 shows two curves obtained from wax pins sliding in the first
case on a clean steel counterface and in the second case on a track of deposited
wax. It is seen that both attain an equilibrium rate of weight loss, but in the
second case the pin shows a distinctly lower rate of wear. It is postulated
that this lower rate of wear is a function of back transfer where the real rate
of wear is masked by the back transferral of material from the deposit.
Measurements showed that the coefficient of friction was higher for the wax
pin sliding on its own deposit. Therefore the difference cannot be attributed
to an increased lubrication effect.
Previous work has shown that weight changes on the counterface are
cyclic in nature in that a regular loss-increase-loss pattern develops. Since
200
180
‘:
P
160
” 120
I
0 loo
;
F 60
’ 60
I
15 30 L5 60 75 90
Shding distance lcml
Fig. 25. Weight loss us. sliding distance for wax pins sliding on a steel counterface : o,
single-track sliding; +, multitrack sliding.
23
the curve of pin weight loss against sliding distance is linear once an equilib-
rium wear rate is attained, this suggests that the bulk of debris generated is
from the counterface itself, but further work is required in this area,
I n viewing the complete range of alloys studied with the various contact
stresses involved it is clear that no single wear mechanism is responsible for
the production of material loss from sliding surfaces. There can in some cases
be a direct effect where material is detached by what has now been accepted
as delegation. However, where material yield stresses are low or contact
stresses are high for a given sliding speed, mutual transfer phenomena suggest
that much of the wear debris accumulated in a wear run is generated from
the action of surface forces on the transferred layers where surface-initiated
cracking reacts with interlayer zones prior to material detachment.
4. Conclusions
(1) I t is postulated that mutual transfer is a significant feature of all
regimes of sliding wear and that it becomes more prevalent in materials whose
yield stresses are low or where normal contact stresses are high.
(2) Asperity interaction, gross layering and ductile fracture constitute
the mode of transfer for conditions bordering on seizure, whereas a shear
transfer mechanism operates in regimes between those of mild and severe
wear.
(3) Delamination producing wear craters is a predominant feature of
mild wear regimes in eutectic and hypereutectic alloys, but primary and
secondary modes, dependent on subsurface and surface failure respectively,
can be recognized.
(4) Delamination of an adhesive variety exists in low yield stress mat-
erials or in materials where high contact stresses exist; this is a function of
surface cracking down to interlayer zones.
References
1 S. Jeharunir, N. P. Suh and E. P. Abrahamson, Weur, 28 (1974) 235.
2 T. Sasada and S. Norose, Froc. Jpn. Sot. Lubr. Eng., (1976) 32.
3 T. Kayaba and K. Kato, Pmt. Int. Co& on the Wear ~~~ateria~, Dearborn, MI , April
16 - 18, 1979, American Society’of Mechanical Engineers, New York, 1979, p. 45.
4 0. Vingsbo, Froc. ht. Conf. on the Wear of Materials, Dearborn, MI , April 16 - 18,
1979, American Society of Mechanical Engineers, New York, 1979, p. 620.
5 E. M. Moore, P. F. Packman and J. J. Wert, ‘I ’ribol. I nt., 7 (1974) 242.
6 D. A. Rigney and W. A. GIaeser, Proc. I nt. Conf. on the Wear of Materiak. St. Louis,
MO, April 25 - 28.1977, American Society of Mechanicel Engineers, New York, 1977,
p. 41.
7 A. A. Torrance, Wear, 50 (1978) 169.
8 N. P. Suh, Wear, 25 (1973) 111.
9 J. Clarke and A. D. Sarkar, Wear, 54 (1979) 7.

Sponsor Documents

Or use your account on DocShare.tips

Hide

Forgot your password?

Or register your new account on DocShare.tips

Hide

Lost your password? Please enter your email address. You will receive a link to create a new password.

Back to log-in

Close