- Email: [email protected]
Dry sliding wear mechanisms in a Ti50Ni47Fe3 intermetallic alloy
WEAR ELSEVIER
Wear
181-183
Dry sliding wear mechanisms
(1995) 302-311
in a T&N&Fe3
intermetallic
alloy
J. Singh, A.T. Alpas Engineering Materi&
Group, Department
of Mechanical
Received
Engineering,
18 May 1994; accepted
University of Windsor, Windsor, Ont. N9B 3P4, Canada 16 September
1994
Abstract T&N&Fe, intermetallic alloy with a B2 structure can be strengthened by means of a thermomechanical treatment to obtain a combination of high strength and high ductility. The strengthened Ti,,Ni,,Fe, alloy with yield strength of 830 MPa and tensile ductility of 15% was tested for dly sliding wear against a SAE 52100 bearing steel using a block on ring wear test configuration. The Ti,,Ni,,Fe, alloy demonstrated excellent wear resistance against SAE 52100 steel in the load range 2-400 N investigated in this study. The wear rates of the T&N&Fe, were found to be 2-5% of that of the SAE 52100 steel despite the lower hardness of the alloy. The metallographic characterization of the worn surfaces indicated that a tribolayer containing mainly iron oxides formed on the wear tracks of the intermetallic. Plastic deformation and mechanical mixing in the subsurface material produced a hardened layer with finger like morphology beneath the oxide layer. The wear resistance of the Ti,,Ni,,Fe, alloy is discussed in terms of the subsurface work hardening and mechanically mixed tribolayer formation on the worn surfaces. Keywords: Sliding wear; Intermetallics;
Ti-Ni alloys; Wear transitions; Thermomechanical
1. Introduction
Titanium-nickel alloys based on the intermetallic compound TiNi are well known for their shape memory properties [1,2]. On cooling, the high temperature ordered phase B2 (CsCl type) undergoes a reversible, diffusionless shear transformation to a distorted monoclinic structure, B19’. Apart from the shape memory effect, TiNi alloys exhibit an unusually high room temperature ductility with respect to other intermetallic compounds, good impact resistance, vibration damping properties as well as good corrosion resistance and biocompatibility [2,3]. Transmission electron microscopy (TEM) analysis of Goo et al. [4] has shown that the B2 phase undergoes twinning during deformation near room temperature. Moberly et al. [S] have further shown that a unique combination of twinning and dislocation slip is responsible for the excellent room temperature ductility of the T&N&Fe, alloy. Their work also showed that it is possible to produce an alloy that combines high strength and ductility through a selective thermomechanical treatment. Melton and Mercier [6] have reported that this alloy work hardens appreciably under cyclic loading and shows an endurance limit close to 0.7 of its ultimate tensile strength. The above properties suggest that TiNi based intermetallic alloys can be good candidates for tribological applications which require wear resistance in addition
0043-1648/95/$09.50 0 1995 Elsevier SSDI 0043-1648(94)07034-2
Science
S.A. All rights
reserved
treatment;
Wear mechanisms; Tribolayers
to good mechanical properties. However, only a few studies have been done to understand the wear characteristics of TiNi base alloys. Suzuki and Kuroyanagi [7] showed that the equiatomic TiNi alloy exhibits very good slurry erosion resistance under corrosive environments. Shida and Sugimoto [S] have reported that the resistance to water jet erosion could be related to the pseudo-elasticity of the TiNi alloy. Wear of a ternary (TiNiFe) alloy similar to the one used in this study has been investigated by Clayton [9] who indicated that the intermetallic alloy exhibits good wear resistance under rolling/sliding contact conditions which can be attributed to the hardening of the alloy under the application of cyclic stresses generated during the sliding wear process. The aim of this study is to investigate the dry sliding wear of the intermetallic Ti,,Ni,,Fe, alloy against a SAE 52100 steel under a wide range of load and sliding speed conditions and to determine the wear mechanisms which operate during sliding wear.
2. Experimental 2.1. Material preparation A ternary alloy of composition T&N&Fe, was used in this study. Substitution of up to 3 at.% Fe for Ni in the binary TiNi alloy lowers the martensitic trans-
J. Singh, A.T. Alpas / Wear 181-183
(1995) 302-311 COLD
formation temperature (MS) to - 140 “C [lo]. The starting materials were titanium sponge (99.9%), nickel pellets (99.9%) and iron lumps (99.98%). Buttons, approximately 50 g in weight, were melted in an arc furnace in an argon atmosphere in a water cooled copper mould. The buttons were turned and remelted 7-8 times to ensure chemical homogeneity. The existence of a single B2 phase in the buttons was confirmed by X-ray diffractometty. The buttons were repeatedly cold worked, with intermediate anneals at 850 “C for 1 h under an argon atmosphere, until the thickness was reduced either to 2 mm sheet for the tensile test specimens or 7 mm for the wear specimens. The optical microscopy revealed that the final grain size of the annealed alloy was about 6 pm (Fig. 1).
303 ROLLED
I
FULL ANNEALING (850°C, 1 hr)
40%
CW
ANNEALING (45O’C.
1
by
WEAR
ROLLING
J
TREATMENT 1 Omin.)
CUT TO (10x10~5
2.2. Tensile testing Tensile testing was done to investigate the effect of the cold working and the thermomechanical treatment on the strength and ductility of the alloy. Ti,&,,Fe, sheets were annealed (850 “C, 1 h) and cold rolled to give different percentages of cold work up to 45% reduction in thickness. Tensile test specimens with a reduced gage section of 0.8 mm thick, 5 mm wide and 20 mm in length were machined from the cold-worked sheets. The tensile tests were done using an Instron Universal Testing machine. A clip-on extensometer fixed on the reduced gage section of the specimens was used for the strain measurements. In a second set of experiments, thermomechanical treatments similar to those used by Moberly et al. [5] were given to the alloy. Cold-worked sheets after 40% reduction in thickness were given a short time anneal at 450 and 500 “C before carrying out tensile tests with the objective of determining the optimum combination of yield strength and ductility.
MATERIAL
SIZS mm
)
1
TESTING
Fig. 2. Outline of the thermomechanical treatment given to T&Ni,,Fe, prior to wear testing.
2.3. Wear testing The dry sliding wear tests were performed using a block-on-ring type test machine. T&N&Fe, alloy was given a thermomechanical treatment as outlined in Fig. 2, consisting of 40% cold work by rolling followed by an isothermal heating at 450 “C for 10 min. Rectangular blocks of 5 x 10x 10 mm3 were machined from the material treated to this condition for wear testing. The reason for this treatment is given in the next section. The rectangular blocks were slid against a SAE type 52100 bearing steel ring (Vickers hardness, WIN, 900 kg mm-‘). The applied load was varied in steps from 2 to 400 N at a constant sliding speed of 0.4 m s-l. In another set of experiments, the sliding speed was varied from 0.04 m s-l to 6 m s-’ while keeping the applied normal load constant at 20 N. The weight of both the block and the ring was measured at regular intervals to determine the weight loss as a function of sliding distance. The worn surfaces were examined visually as well as by optical microscopy. Subsequently a scanning electron microscope (SEM) equipped with an energy dispersive X-ray analysis (EDS) system was used to characterize the worn surfaces. X-ray diffraction analyses (XRD) of wear tracks and wear debris were also performed using a Rigaku X-ray diffractometer with Cu Ka radiation. 3. Results
3.1. Mechanical Fig. 1. Optical micrograph of fully annealed Tis0Ni4,Fe3 intermetallic alloy.
behaviour
The effect of cold work on the properties of the T&N&Fe, alloy is shown in Fig. 3. In the annealed
J. Singh, A.T. Alpas
304
I Wear 181-183
(1995) 302-311
3.2. Wear rates
30 z F:
d 20 2 2
W K
10
0’
I
0
10
I
20 % COLD
I
1
30
40
10 50
WORK
Fig. 3. Effect of cold work on the yield strength tensile strength (UTS), and ductility (percentage Tl,N&Fe,.
(YS), ultimate elongation) of
In Fig. 4(a), the volumetric wear loss of the T&N&Fe, alloy as a result of dry sliding tests conducted at a constant sliding speed of 0.4 m s-l is plotted, against the sliding distance for different applied loads. The wear rates of the intermetallic alloy (slope of the volume loss curves) are high in the beginning but stabilize to lower values at longer sliding distances. This behaviour is observed at all load levels between 2-400 N (this is not self evident in Fig. 4(a) for low load tests due to very low wear rates). The corresponding wear losses of the SAE 52100 steel are shown in Fig. 4(b). These plots differ in an important aspect from those of the Ti,,Ni,,Fe, in that the wear rates do not decrease with sliding distance but continuously increase up to the longest sliding distance used. The average slopes of the volume loss vs. sliding distance curves are plotted against the applied load in Fig. 5. The salient observation I
I
1.5
I
(a)
I
I
I
.
ITable 1 Effect of thermomechanical treatment on the yield strength (YS), ultimate tensile strength (UTS) and ductility (percentage elongation) of T&N&Fe3 Heat treatment after 40% CW by rolling None 450 “C, 450 “C, 450 “C, 500 “C, 500 “C, 850 “C,
5 min 10 min 100 min 5 min 10 min 60 min
1110 1080 830 659 765 708 410
UTS
Elongation
(MPa)
(%)
1450 1290 1100 968 1132 1097 970
4 6 15 17 14 18 20
condition, the alloy is soft with yield strength only 410 MPa and exhibits substantial room temperature ductility (20% elongation). The cold working increases the yield strength appreciably; however, the ductility shows a concomitant decrease. Cold rolling to 45% reduction in thickness increases the yield strength to 1200 MPa and lowers the ductility to 4% elongation. It is clear from Fig. 3 that the intermetallic alloy has not exhausted its work hardening ability after cold working up to 45%. Subsequent anneals between 450-500 “C for short durations were effective in recovering the loss in ductility without adversely effecting the yield strength. These results are in agreement with those reported by Moberly et al. [5]. From Table 1, it is evident that a 10 min anneal at 450 “C gives the best combination of strength and ductility. Hence, this particular treatment was given to the Ti,,Ni,,Fe, alloy before the sliding wear tests.
0
15
10
5 SLIDING
DISTANCE
c
SAE 62100
r
.
0
??
100N 200N 400N
I
I
1
5
10
15
20
SLIDING
DISTANCE
I
0
20
(rnx103)
(mx103)
Fig. 4. Volumetric wear plots of (a) Ti,,Ni,,Fe, alloy and (b) SAJZ 52100 bearing steel at various applied loads at a sliding speed of 0.4 m s-‘.
J. S&h,
T
10-3
A.T. A&as I Wear 181-183
!
,\
E E
z 1o-4
t :
2 3 g
1o-5
j
loo
1 o2
10’ LOAD
Fig. 5. Wear rates of T&N&Fe3 load.
1 o3
(N)
and SAE 52100 steel plotted against
which can be made from these plots is that the wear rates of Ti,,Ni,,Fe, alloy are only about 2 to 5% of those of the SAE 52100. These results are not readily expected since the bearing steel, SAE 52100, is almost three times harder (VHN 900 kg mm-‘) in comparison to the intermetallic alloy (VHN 300 kg mm-‘). The Ti,,Ni,,Fe, alloy shows two distinct wear regimes. At low loads, marked as regime I in Fig. 5, the wear rates are very low (e.g. 1 X 10e6 mm3 m-’ at 2 N and 5 X low6 mm3 m-’ at 9 N) but increase rather rapidly with load. The slope of the wear rate curve decreases with increasing load before reaching a steady value at the transition to regime II. Typical wear rates are 1.5 X 10e5 mm3 m-l at 20 N and 1X 10m4 mm3 m-’ at 200 N in this regime. The slope of the wear rate vs. load curve of the Ti,,Ni,,Fe, remains constant in the entire load range of lo-400 N such that the wear rates, W, in this regime obey a power law expression of the form
305
(1995) 302-311
of the steel increase according to 2.7X 10-5Po.8. Both Ti,,Ni,,Fe, and SAE 52100 show similar values for the wear exponent. However, the wear constant K for the steel is 20 times higher than for the inter-metallic alloy. The effect of sliding speed is investigated for the wear regime II, which dominates the wear behaviour of the Ti5,,Ni_,7Fe3aswell as the SAE 52100. The results of these tests carried out at 20 N load are shown in Fig. 6. Sharp transitions in the wear rates of both materials are self evident. At low sliding speeds, from 0.04 to 2.7 m s-l, the wear rates of the T&Ni4,Fe3 are very low (characteristic of regime II of Fig. 5) and increase gradually with the sliding speed (e.g. 1.5 x 10e5 mm3 m-l at 0.4 m s-l to 2.0X lo-’ mm3 m-’ at 2.7 m s-l). However, the wear rates rise sharply by a factor of 10 (to 2~10~~ mm3 m-‘) when the sliding speed is increased to 3.0 m SC’. This mild to severe wear transition of the Ti5ai4,Fe3 occurs within a very narrow range of sliding speed (2.7-3.2 m s-l). Further increase in the sliding speed increases the wear rates only slightly (to 2.5~10~~ mm3 m-l at 5.3 m s-l). The SAE 52100 bearing steel, in contrast to Ti,,Ni,,Fe,, shows a sharp fall in the wear rate at a critical sliding speed of 2.5 m s-l. The wear rates of the steel drop by a factor of 20 from 6~ 10m4 mm3 m -’ to 3X10m5 mm3 m-’ on increasing the sliding speed above 2.5 m s-l. Moreover, the wear rates of the steel continue to decrease as the sliding speed is increased beyond this transition speed. It is important to note that the wear transition in the SAE 52100 (high to low wear rates) occurs at a lower sliding velocity with respect to the mild to severe wear transition speed of the Ti,,Ni,,Fe,. Thus, the high wear rates of the
1 o-3
\+z “E E
W=kP” where P is applied normal load, and K and it are defined as the wear constant and the wear exponent, respectively. The wear rate curve for the steel has a shape similar to that of Ti5ai4,Fe3 except for high loads (>200 N), at which the wear rates accelerate and, thus, show another transition which is labelled as regime III in Fig. 5. The wear data of the Ti,$Ii,,Fe, as well as the SAE 52100 for regime II fitted to the above expression show that the wear rates of the intermetallic vary as 1.4 X 10w6Po.’ while the wear rates
1o-5
1 0
I
I
I
1
2
3
SLIDING
SPEED
I
I
4 5 (m/s)
6
I
Fig. 6. Wear rates of T&N&Fe3 and SAE 52100 plotted against sliding speed (applied load is 20 N).
.I. Singh, A. T. Alpas I Wear 181-183
306
steel are not a precondition Ti,,Ni,,Fe,.
(1995) 302-311
for the mild wear of the
3.3. Worn surfaces and wear debris The visual examination of the worn surfaces of Ti,,Ni,,Fe, blocks tested in regime I and II revealed the presence of reddish-brown layers on the wear surfaces along with continuous groves in the direction of sliding. The low power optical microscopy (Fig. 7) showed that the wear surfaces are covered with finely compacted powder. Energy dispersive X-ray (EDS) analysis (Fig. 8) showed that an iron rich subsurface layer developed on the wear track. The amount of iron
I
0
1
2
3
X-RAY
1
I
4
5
ENERGY
I
6
7
6
9
10
(KeV)
Fig. 8. Results of the energy dispersive X-ray analysis performed on the wear tracks after sliding against SAE 52100 at (a) 10 N load and (b) 400 N load (sliding speed 0.4 m s-l).
on the surface increased with load and sliding distance. At high loads, the layer became discontinuous and smeared along the groves on the worn surface. These layers at high loads appeared to be black in colour to the unaided eye. The wear debris produced during wear tests was examined by metallography as well as X-ray diffraction analysis (XRD), and found to vary markedly in colour and composition with the applied load. The XRD results for the debris produced in the different wear regimes are shown in Fig. 9 Four types of wear debris were observed: (1) Reddish fine powder, mainly Fe,O, (in wear regime I); (2) Dark brown, fine powder containing Fe,O, and Fe,O, (in wear regime II); (3) Fine black powder mixed with large metallic debris having metallic iron, Fe,O, as well as Fe,O, (in wear regime III); (4) Large metallic (= 100 pm), Ti,,Ni,,Fe, plates, at sliding speeds above 3.0 m s-‘. The worn surface of the steel ring was found to be shiny except for the formation of continuous groves in the sliding direction. The electron micrograph (Fig. 10) shows that the surface within these groves has a polished appearance. EDS results did not show evidence for the material transfer from T&N&Fe, to the steel. 3.4. Subsurface microstmcture
Fig. 7. Worn surfaces of Ti,,Ni,,Fe, at 0.4 m SC’ sliding speed: (a) optical micrograph, 10 N load; (b) scanning electron micrograph, 10 N load; (c) optical micrograph, 400 N load.
The optical micrograph (Fig. 11) of a cross-section parallel to the sliding direction shows that extensive plastic deformation took place below the contact surface during sliding. In a deformation zone of about 100 pm deep, the grains are elongated towards the sliding direction. The worn specimen cut and polished at a taper angle of 1-2” to the surface was also investigated.
J. Singh, A.T. Alpas / Wear 181483
20
10
30
40
50
60
50
60
(199.5) 302-311
70
307
60
90
100
(b)
28 Fig. 9. X-ray diffraction analysis (XRD)
(DEGREES)
of the debris at (a) 5 N load and (b) 400 N load (sliding speed 0.4 m s-l).
A “finger like” morphology, consisting of features elongated in the direction of sliding developed right below the worn surface as shown in Fig. 12(a). These “fingers” form directly below the groves on the worn surface and were found to penetrate to a depth of about SO-100 pm below the surface. According to the EDS analysis they are composed of a mixture of Ti, Ni and Fe. Iron consistently exhibited the strongest intensity throughout the layer. Fig. 13 shows dark veins running parallel to the surface at a depth of about 40 pm within the mechanically mixed layer. The veins were observed to be made up of iron and presumably of iron oxides (Fe,O, and Fe,O,). Microhardness (Vickers) measurements taken on the taper sections showed that the mechanically mixed layers (“fingers”) have hardness values ranging from 500 to 800 kg mm-’ as compared to 350-400 kg mrr-’ for the material outside them (Fig. 12(b)). EDS analysis
of the cross-section of a mechanically mixed layer is shown in Fig. 13. The fingers were found to be resistant to abrasive wear. Metallographic polishing on wet cloth with 0.05 pm Al,O, powder removed the material from the surrounding region at a faster rate than the fingers and therefore the fingers appear as elevated plateaus. Monitoring the size of the microhardness indentations before and after the polishing also confirmed that the material removal rate from the fingers was much lower. In summary, the results indicate that the worn surfaces of Ti,,Ni,,Fe, are severely deformed during sliding. The surfaces were covered by an iron rich layer which consisted of compacted oxide particles. Metallography indicated that beneath this layer are finger like features which are mechanically mixed tribolayers composed of iron (possibly Fe,O, and Fe,O,) and Ti and Ni compounds. The wear characteristics of the intermetallic
308
J. Singh, A.T. Alpas
I Wear 181-183
(1995) 302-311
Fig. 10. Surface morphology of SAE 521000 steel ring after wear test at 400 N load (sliding speed 0.4 m SK’).
Fig. 11. Longitudinal cross-section through the worn surface revealing the extent of plastic deformation during wear of the T&,Ni,,Fe, at 400 N load (sliding speed 0.4 m SC’).
and the steel as a function of load and sliding velocity are summarized in Tables 2 and 3, respectively.
4. Discussion One of the most important results obtained in the present study is that the intermetallic Ti,,Ni,,Fe, shows 2-5% as much wear as steel during dry sliding wear against the SAE 52100 steel although the intermetallic has a lower bulk hardness (300 kg mm-‘) in comparison to that of the SAE 52100 bearing steel (900 kg mm-‘). EDS results (Fig. 8) as well as worn surface metallography (Fig. 7) clearly reveal that a layer containing mainly iron and its oxides forms on the wear surface during the wear process. The presence of iron oxides lowers the wear damage by acting as solid lubricants as well as reducing the metal to metal contact [11,12]. In the beginning of the wear process, Ti,,Ni,,Fe, makes
Fig. 12. Tapered section of the intermetallic block tested under 400 N load (l-2” to the worn surface), depicting (a) finger type morphology below the contact surface and (b) microhardness values obtained at various locations at the surface in (a) (sliding speed 0.4 m SC’).
direct contact with the steel ring and therefore shows high wear rates (Fig. 4(a)). When a layer of iron oxides starts to form on the surface of the Ti,,Ni,,Fe,, the wear rates show a decrease with increasing sliding distance until this layer reaches an equilibrium thickness to give steady state wear rates. The intermetallic shows in general a high oxidation resistance and therefore can be expected to show lower oxidative wear rates than the steel counterface. In ductile materials wear processes generate large plastic strains beneath the contact surfaces. It is shown by cold rolling experiments (Fig. 3) that Ti,,Ni,,Fe, can considerably work harden at high strains. In fact, the microhardness measurements taken on the contact surface, in locations outside the mechanically mixed finger like features, indicate that the surface hardness of T&N&Fe, can reach up to 400 kg mm-‘, an increase of about 30% relative to the bulk hardness. Thus, the hardening of the contact surfaces during wear may increase the wear resistance of Ti,,Ni,,Fe,. It is im-
J. Singh, A.T. A&as
/ Wear 181-183
(1995) 302-311
0
i
2
3
309
4
X-RAY
5
6
ENERGY
7
a
s
B
BlO
8
9
10
(KsV)
I”“““’
j 0
12
3
4
X-RAY
2
3 X-RAY
4
5
6
ENERGY
7
B
a
10
0
1
2
(KeV)
3 X-RAY
5
6
ENERGY
4
5
7
(KeV)
6
ENERGY
7
10
(KsV)
Fig. 13. Transverse cross-section through a mechanicahy mixed layer (finger type morphology) along with EDS analysis results. Locations where the analysis was performed arc indicated by letters a-d. Table 2 Summary of wear characteristics
as a function of load Regime I
Regime II
Regime III
1 x 10-5-1 x 10-d 2x 10-4-2 x lo-’
-
Wear rates (mm’ m-‘)
Ti,$Ji,,Fe3 ME52100
1 x10-6-1 x10-5 2x1o-5-2x1o-4
Worn surface
Tis,,Ni4,Fe3
??
SAE 52100 Wear debris
smooth ?? reddish brown (oxide) layer ?? no plastic deformation ??
smooth and shiny
fine powder reddish brown ?? oxide (FerO3)
shallow groves 0 dark brown (oxide) layer ?? plastic deformation ?? hardened layer ?? fine groves, shiny ??
fine powder dark brown 0 oxide (Fe203,Fe304)
>2x10-3 rough with deep groves black layer (Fe rich) ?? heavy plastic deformation ?? formation of hardened fingers ?? uneven shiny ??
??
black and mixed large metallic Fe and ?? fine oxide (FezOp and Fea04)
??
??
??
??
??
??
portant to note that the cold worked inter-metallic material can retain its strength at elevated temperatures (Table l), so that, once formed, the hard surface layers are not expected to soften while the contact surface temperature increases during the sliding process. High hardness values (Fig. 12(b)) recorded in the mechanically mixed layers may be attributed to a structure which consists of a mixture of iron oxide debris with the base material. Formation of mechanically mixed layers on the worn surfaces have been reported by
several researchers [13,14]. TEM examination by Rainforth et al. [14] showed that the submicroscopic iron oxide particles penetrate up to a depth of several tens of microns and intimately mix with the base material. This is in agreement with the observation that the layers of a mechanical mixture of iron oxides and intermetallic compounds have penetrated into the T&N&Fe, up to a depth of about 80 pm. Although, details of the formation of iron oxide rich finger like features are not clear, their formation can be the result
310
.I. Singh, A.T. Alps
Table 3 Summary
of wear characteristics
as a function
of sliding
Sliding
speed
/ Wear 181-183
speed <2.5
m SC’ (pre-transition)
Wear rates (mm’ m-‘)
TiSONi,,Fe, SAE. 52100
1 x 10-5-2x 3x 1o-4-6x
Worn
Ti,,Ni,,Fe,
?? reddish brown 0 fine groves ?? subsurface deformation ?? hardened layer
Wear
surface
debris
(1995) 302-311
-
speed
>3.0
m SK’ (post-transition)
2x10-4-3x10-4 3x10-5-2x10-5
10-S 1o-4
black patches rough with deep groves ?? minimal deformation ?? subsurface cracks ??
??
0 fine powder reddish brown ?? iron oxides (Fe203,Fe304)
??
??
of the penetration of the mechanically mixed material into the subsurface material. Don et al. [15] have indicated that, whenever the surface hardness H is greater than the hardness h of the base material, i.e. H/h > 1.2, the hardened layer will penetrate and sink into the subsurface region. The wear of T&N&Fe, can be described as mild at low sliding speeds and severe at high sliding speeds (greater than 3.0 m s-l) as indicated by the shape of the wear rate vs. the sliding velocity plot of Fig. 6. At low sliding speeds, the low wear rates of this alloy can be attributed to the work hardening of the surface layers and the formation of a hardened wear resistant layer which consists of iron oxides and the base alloy. However, above a critical sliding speed of about 3.0 m s-‘, the protective tribolayers cease to form. Metallographic examination of the subsurface region showed no sign of formation of finger like features or iron oxide veins at high sliding speeds. Instead, cracks emanating from the bottom of the surface groves running normal to the wear surface were observed. The wear process, under these conditions, produces metallic debris in the 100 pm size range. It appears that the ductility of T&N&Fe, is adversely effected by the high strain rates caused by high sliding speeds. Subsurface crack formation could also be the result of thermal fatigue due to higher temperature gradients and thermal cycling of the surface which is no longer protected by the tribolayers. The possibility of the martensitic transformation has also been considered as an alternative explanation of the transition to severe wear since the martensite of T&N&Fe3 is brittle and much softer [16,17]. XRD analysis performed on the debris generated at 4 m s-l did not reveal martensite. This was nevertheless expected since martensite, if it could be induced during sliding, would be pseudo-elastic and would transform back to the austenite phase once the stress is removed. However, formation of stress induced martensite during sliding is unlikely. Firstly, the MS temperature of T&N&Fe, is - 140 “C [lo], whereas, the temperature near the contact surface (at a sliding speed of 4 m
Sliding
??
large (= 100 wrn) metallic (Ti5aNt4,Fe3)
s-l and 20 N load) was measured to be 170 “C only after 10 min (by inserting a 0.2 mm thick ungrounded thermocouple about 50 pm below the contact surface). Subsurface temperature increases with speed and the stresses needed for the stress induced martensite can easily reach 3-5 GPa [16,18]. Secondly, mechanical twinning rather than martensitic transformation is preferred at ambient and high temperatures [5]. The wear of the SAF 52100 steel undergoes a change in wear mechanism at a sliding speed of 2.6 m s-l. At this sliding speed, the steel appears to undergo a transition from mild oxidative type wear to severe oxidative wear, resulting in a large drop in its wear rates. Such a transition has been related to the formation of an oxide layer on the steel surface itself when the temperature of the steel is sufficiently high to reform oxide at a rate faster than it is removed. Increase in the sliding speed increases the surface temperature and therefore the oxidation rate of iron.
5. Conclusions 1.
2.
3.
A thermomechanical treatment (45% cold work followed by annealing at 450 “C for 10 min) can be used to strengthen T&N&Fe, intermetallic alloy to obtain an optimum combination of the yield strength (830 MPa) and ductility (15% tensile elongation). The strengthened T&N&Fe, alloy shows excellent wear resistance with wear rates between 10m6 and 10e5 mm3 m-’ in the load range 2-400 N, when tested for dry sliding wear against a SAE 52100 steel. These wear rates of the T&N&Fe, are only 2-5% of those of the SAE 52100 bearing steel counterface. During sliding wear, surface layers of the Ti,,Ni,,Fe, work harden and also form finger like features with very high hardness. These layers, containing a mechanical mixture of iron oxides and constituents of the base material, form beneath the contact surface. The surfaces themselves are covered with Fe,O,
J. Singh, A.T. Alpas / Wear 181-183
4.
and Fe,O, oxidized wear products transferred from the counterface. At high sliding speeds, higher than 3.0 m s-l, T&N&Fe, suffers severe wear. However, the steel counterface shows a transition from mild oxidative wear to severe oxidative wear when the sliding speed exceeds 2.6 m s-l.
Acknowledgements The authors gratefully acknowledge financial support provided by Natural Science and Engineering Research Council of Canada (NSERC).
References [l] [2]
S.P. Gupta and A.A. Johnson, Trans. Jpn. Inst. Met., 14 (1973) 292. C.M. Wayman, J. Met, 6 (1980) 129.
(1995) 302-311
311
and R.J. Biermann, J. Met., 2 (1988) 32. [31 SM. Tuominen [41 E. Goo, T. Duerig, K. Melton and R. Sinclair, Acta MetaN., 33 (1985) 1725. [51 W.J. Moberly, J.L. Proft, T.W. Duerig and R. Sinclair, Acta Metall. Mater., 38 (1990) 2601. Fl K.N. Melton and 0. Mercier, Acta Metall., 27 (1979) 137. Titanium Zirconium, 27 (1979) [71 Y. Suzuki and T. Kuroyanagi, 67. Wear, 146 (1991) 219. [81 Y. Shida and Y. Sugimoto, [91 P. Clayton, Wear, 162-164 (1993) 202. [lOI E. Goo and R. Sinclair, Acfa Metall., 33 (1985) 1717. T.F.J. Quinn, Br. L Appl. Phys., 13 (1962) 33. PI P21 H. Kato, T.S. Eyre and B. Ralph, Acta Metall. Mater., 42 (1994) 1703. P31 P. Heilmann, J. Don, T.C. Sun, D.A. Rigney and W.A. Glaeser, Wear, 91 (1983) 171. R. Stevens and J. Nutting, Phi/ox Msg., A66 1141 W.M. Rainforth, (1992) 621. 1151 J. Don, T.C. Sun and D.A. Rigney, Wear, 91 (1983) 191. J. Perkins and J.M. Johnson, Ser. Metall., 9 1161 G.R. Edwards, (1975) 1167. P71 S. Miyazaki, T. Imai, Y. Igo and K. Otsuka, Metall. Trans., 17A (1986) 115. P81 F. Takei, T. Miura, S. Miyazaki, S. Kimura, K. Otsuka and Y. Suzuki, Ser. Metall., 17 (1983) 987.
Wear
181-183
Dry sliding wear mechanisms
(1995) 302-311
in a T&N&Fe3
intermetallic
alloy
J. Singh, A.T. Alpas Engineering Materi&
Group, Department
of Mechanical
Received
Engineering,
18 May 1994; accepted
University of Windsor, Windsor, Ont. N9B 3P4, Canada 16 September
1994
Abstract T&N&Fe, intermetallic alloy with a B2 structure can be strengthened by means of a thermomechanical treatment to obtain a combination of high strength and high ductility. The strengthened Ti,,Ni,,Fe, alloy with yield strength of 830 MPa and tensile ductility of 15% was tested for dly sliding wear against a SAE 52100 bearing steel using a block on ring wear test configuration. The Ti,,Ni,,Fe, alloy demonstrated excellent wear resistance against SAE 52100 steel in the load range 2-400 N investigated in this study. The wear rates of the T&N&Fe, were found to be 2-5% of that of the SAE 52100 steel despite the lower hardness of the alloy. The metallographic characterization of the worn surfaces indicated that a tribolayer containing mainly iron oxides formed on the wear tracks of the intermetallic. Plastic deformation and mechanical mixing in the subsurface material produced a hardened layer with finger like morphology beneath the oxide layer. The wear resistance of the Ti,,Ni,,Fe, alloy is discussed in terms of the subsurface work hardening and mechanically mixed tribolayer formation on the worn surfaces. Keywords: Sliding wear; Intermetallics;
Ti-Ni alloys; Wear transitions; Thermomechanical
1. Introduction
Titanium-nickel alloys based on the intermetallic compound TiNi are well known for their shape memory properties [1,2]. On cooling, the high temperature ordered phase B2 (CsCl type) undergoes a reversible, diffusionless shear transformation to a distorted monoclinic structure, B19’. Apart from the shape memory effect, TiNi alloys exhibit an unusually high room temperature ductility with respect to other intermetallic compounds, good impact resistance, vibration damping properties as well as good corrosion resistance and biocompatibility [2,3]. Transmission electron microscopy (TEM) analysis of Goo et al. [4] has shown that the B2 phase undergoes twinning during deformation near room temperature. Moberly et al. [S] have further shown that a unique combination of twinning and dislocation slip is responsible for the excellent room temperature ductility of the T&N&Fe, alloy. Their work also showed that it is possible to produce an alloy that combines high strength and ductility through a selective thermomechanical treatment. Melton and Mercier [6] have reported that this alloy work hardens appreciably under cyclic loading and shows an endurance limit close to 0.7 of its ultimate tensile strength. The above properties suggest that TiNi based intermetallic alloys can be good candidates for tribological applications which require wear resistance in addition
0043-1648/95/$09.50 0 1995 Elsevier SSDI 0043-1648(94)07034-2
Science
S.A. All rights
reserved
treatment;
Wear mechanisms; Tribolayers
to good mechanical properties. However, only a few studies have been done to understand the wear characteristics of TiNi base alloys. Suzuki and Kuroyanagi [7] showed that the equiatomic TiNi alloy exhibits very good slurry erosion resistance under corrosive environments. Shida and Sugimoto [S] have reported that the resistance to water jet erosion could be related to the pseudo-elasticity of the TiNi alloy. Wear of a ternary (TiNiFe) alloy similar to the one used in this study has been investigated by Clayton [9] who indicated that the intermetallic alloy exhibits good wear resistance under rolling/sliding contact conditions which can be attributed to the hardening of the alloy under the application of cyclic stresses generated during the sliding wear process. The aim of this study is to investigate the dry sliding wear of the intermetallic Ti,,Ni,,Fe, alloy against a SAE 52100 steel under a wide range of load and sliding speed conditions and to determine the wear mechanisms which operate during sliding wear.
2. Experimental 2.1. Material preparation A ternary alloy of composition T&N&Fe, was used in this study. Substitution of up to 3 at.% Fe for Ni in the binary TiNi alloy lowers the martensitic trans-
J. Singh, A.T. Alpas / Wear 181-183
(1995) 302-311 COLD
formation temperature (MS) to - 140 “C [lo]. The starting materials were titanium sponge (99.9%), nickel pellets (99.9%) and iron lumps (99.98%). Buttons, approximately 50 g in weight, were melted in an arc furnace in an argon atmosphere in a water cooled copper mould. The buttons were turned and remelted 7-8 times to ensure chemical homogeneity. The existence of a single B2 phase in the buttons was confirmed by X-ray diffractometty. The buttons were repeatedly cold worked, with intermediate anneals at 850 “C for 1 h under an argon atmosphere, until the thickness was reduced either to 2 mm sheet for the tensile test specimens or 7 mm for the wear specimens. The optical microscopy revealed that the final grain size of the annealed alloy was about 6 pm (Fig. 1).
303 ROLLED
I
FULL ANNEALING (850°C, 1 hr)
40%
CW
ANNEALING (45O’C.
1
by
WEAR
ROLLING
J
TREATMENT 1 Omin.)
CUT TO (10x10~5
2.2. Tensile testing Tensile testing was done to investigate the effect of the cold working and the thermomechanical treatment on the strength and ductility of the alloy. Ti,&,,Fe, sheets were annealed (850 “C, 1 h) and cold rolled to give different percentages of cold work up to 45% reduction in thickness. Tensile test specimens with a reduced gage section of 0.8 mm thick, 5 mm wide and 20 mm in length were machined from the cold-worked sheets. The tensile tests were done using an Instron Universal Testing machine. A clip-on extensometer fixed on the reduced gage section of the specimens was used for the strain measurements. In a second set of experiments, thermomechanical treatments similar to those used by Moberly et al. [5] were given to the alloy. Cold-worked sheets after 40% reduction in thickness were given a short time anneal at 450 and 500 “C before carrying out tensile tests with the objective of determining the optimum combination of yield strength and ductility.
MATERIAL
SIZS mm
)
1
TESTING
Fig. 2. Outline of the thermomechanical treatment given to T&Ni,,Fe, prior to wear testing.
2.3. Wear testing The dry sliding wear tests were performed using a block-on-ring type test machine. T&N&Fe, alloy was given a thermomechanical treatment as outlined in Fig. 2, consisting of 40% cold work by rolling followed by an isothermal heating at 450 “C for 10 min. Rectangular blocks of 5 x 10x 10 mm3 were machined from the material treated to this condition for wear testing. The reason for this treatment is given in the next section. The rectangular blocks were slid against a SAE type 52100 bearing steel ring (Vickers hardness, WIN, 900 kg mm-‘). The applied load was varied in steps from 2 to 400 N at a constant sliding speed of 0.4 m s-l. In another set of experiments, the sliding speed was varied from 0.04 m s-l to 6 m s-’ while keeping the applied normal load constant at 20 N. The weight of both the block and the ring was measured at regular intervals to determine the weight loss as a function of sliding distance. The worn surfaces were examined visually as well as by optical microscopy. Subsequently a scanning electron microscope (SEM) equipped with an energy dispersive X-ray analysis (EDS) system was used to characterize the worn surfaces. X-ray diffraction analyses (XRD) of wear tracks and wear debris were also performed using a Rigaku X-ray diffractometer with Cu Ka radiation. 3. Results
3.1. Mechanical Fig. 1. Optical micrograph of fully annealed Tis0Ni4,Fe3 intermetallic alloy.
behaviour
The effect of cold work on the properties of the T&N&Fe, alloy is shown in Fig. 3. In the annealed
J. Singh, A.T. Alpas
304
I Wear 181-183
(1995) 302-311
3.2. Wear rates
30 z F:
d 20 2 2
W K
10
0’
I
0
10
I
20 % COLD
I
1
30
40
10 50
WORK
Fig. 3. Effect of cold work on the yield strength tensile strength (UTS), and ductility (percentage Tl,N&Fe,.
(YS), ultimate elongation) of
In Fig. 4(a), the volumetric wear loss of the T&N&Fe, alloy as a result of dry sliding tests conducted at a constant sliding speed of 0.4 m s-l is plotted, against the sliding distance for different applied loads. The wear rates of the intermetallic alloy (slope of the volume loss curves) are high in the beginning but stabilize to lower values at longer sliding distances. This behaviour is observed at all load levels between 2-400 N (this is not self evident in Fig. 4(a) for low load tests due to very low wear rates). The corresponding wear losses of the SAE 52100 steel are shown in Fig. 4(b). These plots differ in an important aspect from those of the Ti,,Ni,,Fe, in that the wear rates do not decrease with sliding distance but continuously increase up to the longest sliding distance used. The average slopes of the volume loss vs. sliding distance curves are plotted against the applied load in Fig. 5. The salient observation I
I
1.5
I
(a)
I
I
I
.
ITable 1 Effect of thermomechanical treatment on the yield strength (YS), ultimate tensile strength (UTS) and ductility (percentage elongation) of T&N&Fe3 Heat treatment after 40% CW by rolling None 450 “C, 450 “C, 450 “C, 500 “C, 500 “C, 850 “C,
5 min 10 min 100 min 5 min 10 min 60 min
1110 1080 830 659 765 708 410
UTS
Elongation
(MPa)
(%)
1450 1290 1100 968 1132 1097 970
4 6 15 17 14 18 20
condition, the alloy is soft with yield strength only 410 MPa and exhibits substantial room temperature ductility (20% elongation). The cold working increases the yield strength appreciably; however, the ductility shows a concomitant decrease. Cold rolling to 45% reduction in thickness increases the yield strength to 1200 MPa and lowers the ductility to 4% elongation. It is clear from Fig. 3 that the intermetallic alloy has not exhausted its work hardening ability after cold working up to 45%. Subsequent anneals between 450-500 “C for short durations were effective in recovering the loss in ductility without adversely effecting the yield strength. These results are in agreement with those reported by Moberly et al. [5]. From Table 1, it is evident that a 10 min anneal at 450 “C gives the best combination of strength and ductility. Hence, this particular treatment was given to the Ti,,Ni,,Fe, alloy before the sliding wear tests.
0
15
10
5 SLIDING
DISTANCE
c
SAE 62100
r
.
0
??
100N 200N 400N
I
I
1
5
10
15
20
SLIDING
DISTANCE
I
0
20
(rnx103)
(mx103)
Fig. 4. Volumetric wear plots of (a) Ti,,Ni,,Fe, alloy and (b) SAJZ 52100 bearing steel at various applied loads at a sliding speed of 0.4 m s-‘.
J. S&h,
T
10-3
A.T. A&as I Wear 181-183
!
,\
E E
z 1o-4
t :
2 3 g
1o-5
j
loo
1 o2
10’ LOAD
Fig. 5. Wear rates of T&N&Fe3 load.
1 o3
(N)
and SAE 52100 steel plotted against
which can be made from these plots is that the wear rates of Ti,,Ni,,Fe, alloy are only about 2 to 5% of those of the SAE 52100. These results are not readily expected since the bearing steel, SAE 52100, is almost three times harder (VHN 900 kg mm-‘) in comparison to the intermetallic alloy (VHN 300 kg mm-‘). The Ti,,Ni,,Fe, alloy shows two distinct wear regimes. At low loads, marked as regime I in Fig. 5, the wear rates are very low (e.g. 1 X 10e6 mm3 m-’ at 2 N and 5 X low6 mm3 m-’ at 9 N) but increase rather rapidly with load. The slope of the wear rate curve decreases with increasing load before reaching a steady value at the transition to regime II. Typical wear rates are 1.5 X 10e5 mm3 m-l at 20 N and 1X 10m4 mm3 m-’ at 200 N in this regime. The slope of the wear rate vs. load curve of the Ti,,Ni,,Fe, remains constant in the entire load range of lo-400 N such that the wear rates, W, in this regime obey a power law expression of the form
305
(1995) 302-311
of the steel increase according to 2.7X 10-5Po.8. Both Ti,,Ni,,Fe, and SAE 52100 show similar values for the wear exponent. However, the wear constant K for the steel is 20 times higher than for the inter-metallic alloy. The effect of sliding speed is investigated for the wear regime II, which dominates the wear behaviour of the Ti5,,Ni_,7Fe3aswell as the SAE 52100. The results of these tests carried out at 20 N load are shown in Fig. 6. Sharp transitions in the wear rates of both materials are self evident. At low sliding speeds, from 0.04 to 2.7 m s-l, the wear rates of the T&Ni4,Fe3 are very low (characteristic of regime II of Fig. 5) and increase gradually with the sliding speed (e.g. 1.5 x 10e5 mm3 m-l at 0.4 m s-l to 2.0X lo-’ mm3 m-’ at 2.7 m s-l). However, the wear rates rise sharply by a factor of 10 (to 2~10~~ mm3 m-‘) when the sliding speed is increased to 3.0 m SC’. This mild to severe wear transition of the Ti5ai4,Fe3 occurs within a very narrow range of sliding speed (2.7-3.2 m s-l). Further increase in the sliding speed increases the wear rates only slightly (to 2.5~10~~ mm3 m-l at 5.3 m s-l). The SAE 52100 bearing steel, in contrast to Ti,,Ni,,Fe,, shows a sharp fall in the wear rate at a critical sliding speed of 2.5 m s-l. The wear rates of the steel drop by a factor of 20 from 6~ 10m4 mm3 m -’ to 3X10m5 mm3 m-’ on increasing the sliding speed above 2.5 m s-l. Moreover, the wear rates of the steel continue to decrease as the sliding speed is increased beyond this transition speed. It is important to note that the wear transition in the SAE 52100 (high to low wear rates) occurs at a lower sliding velocity with respect to the mild to severe wear transition speed of the Ti,,Ni,,Fe,. Thus, the high wear rates of the
1 o-3
\+z “E E
W=kP” where P is applied normal load, and K and it are defined as the wear constant and the wear exponent, respectively. The wear rate curve for the steel has a shape similar to that of Ti5ai4,Fe3 except for high loads (>200 N), at which the wear rates accelerate and, thus, show another transition which is labelled as regime III in Fig. 5. The wear data of the Ti,$Ii,,Fe, as well as the SAE 52100 for regime II fitted to the above expression show that the wear rates of the intermetallic vary as 1.4 X 10w6Po.’ while the wear rates
1o-5
1 0
I
I
I
1
2
3
SLIDING
SPEED
I
I
4 5 (m/s)
6
I
Fig. 6. Wear rates of T&N&Fe3 and SAE 52100 plotted against sliding speed (applied load is 20 N).
.I. Singh, A. T. Alpas I Wear 181-183
306
steel are not a precondition Ti,,Ni,,Fe,.
(1995) 302-311
for the mild wear of the
3.3. Worn surfaces and wear debris The visual examination of the worn surfaces of Ti,,Ni,,Fe, blocks tested in regime I and II revealed the presence of reddish-brown layers on the wear surfaces along with continuous groves in the direction of sliding. The low power optical microscopy (Fig. 7) showed that the wear surfaces are covered with finely compacted powder. Energy dispersive X-ray (EDS) analysis (Fig. 8) showed that an iron rich subsurface layer developed on the wear track. The amount of iron
I
0
1
2
3
X-RAY
1
I
4
5
ENERGY
I
6
7
6
9
10
(KeV)
Fig. 8. Results of the energy dispersive X-ray analysis performed on the wear tracks after sliding against SAE 52100 at (a) 10 N load and (b) 400 N load (sliding speed 0.4 m s-l).
on the surface increased with load and sliding distance. At high loads, the layer became discontinuous and smeared along the groves on the worn surface. These layers at high loads appeared to be black in colour to the unaided eye. The wear debris produced during wear tests was examined by metallography as well as X-ray diffraction analysis (XRD), and found to vary markedly in colour and composition with the applied load. The XRD results for the debris produced in the different wear regimes are shown in Fig. 9 Four types of wear debris were observed: (1) Reddish fine powder, mainly Fe,O, (in wear regime I); (2) Dark brown, fine powder containing Fe,O, and Fe,O, (in wear regime II); (3) Fine black powder mixed with large metallic debris having metallic iron, Fe,O, as well as Fe,O, (in wear regime III); (4) Large metallic (= 100 pm), Ti,,Ni,,Fe, plates, at sliding speeds above 3.0 m s-‘. The worn surface of the steel ring was found to be shiny except for the formation of continuous groves in the sliding direction. The electron micrograph (Fig. 10) shows that the surface within these groves has a polished appearance. EDS results did not show evidence for the material transfer from T&N&Fe, to the steel. 3.4. Subsurface microstmcture
Fig. 7. Worn surfaces of Ti,,Ni,,Fe, at 0.4 m SC’ sliding speed: (a) optical micrograph, 10 N load; (b) scanning electron micrograph, 10 N load; (c) optical micrograph, 400 N load.
The optical micrograph (Fig. 11) of a cross-section parallel to the sliding direction shows that extensive plastic deformation took place below the contact surface during sliding. In a deformation zone of about 100 pm deep, the grains are elongated towards the sliding direction. The worn specimen cut and polished at a taper angle of 1-2” to the surface was also investigated.
J. Singh, A.T. Alpas / Wear 181483
20
10
30
40
50
60
50
60
(199.5) 302-311
70
307
60
90
100
(b)
28 Fig. 9. X-ray diffraction analysis (XRD)
(DEGREES)
of the debris at (a) 5 N load and (b) 400 N load (sliding speed 0.4 m s-l).
A “finger like” morphology, consisting of features elongated in the direction of sliding developed right below the worn surface as shown in Fig. 12(a). These “fingers” form directly below the groves on the worn surface and were found to penetrate to a depth of about SO-100 pm below the surface. According to the EDS analysis they are composed of a mixture of Ti, Ni and Fe. Iron consistently exhibited the strongest intensity throughout the layer. Fig. 13 shows dark veins running parallel to the surface at a depth of about 40 pm within the mechanically mixed layer. The veins were observed to be made up of iron and presumably of iron oxides (Fe,O, and Fe,O,). Microhardness (Vickers) measurements taken on the taper sections showed that the mechanically mixed layers (“fingers”) have hardness values ranging from 500 to 800 kg mm-’ as compared to 350-400 kg mrr-’ for the material outside them (Fig. 12(b)). EDS analysis
of the cross-section of a mechanically mixed layer is shown in Fig. 13. The fingers were found to be resistant to abrasive wear. Metallographic polishing on wet cloth with 0.05 pm Al,O, powder removed the material from the surrounding region at a faster rate than the fingers and therefore the fingers appear as elevated plateaus. Monitoring the size of the microhardness indentations before and after the polishing also confirmed that the material removal rate from the fingers was much lower. In summary, the results indicate that the worn surfaces of Ti,,Ni,,Fe, are severely deformed during sliding. The surfaces were covered by an iron rich layer which consisted of compacted oxide particles. Metallography indicated that beneath this layer are finger like features which are mechanically mixed tribolayers composed of iron (possibly Fe,O, and Fe,O,) and Ti and Ni compounds. The wear characteristics of the intermetallic
308
J. Singh, A.T. Alpas
I Wear 181-183
(1995) 302-311
Fig. 10. Surface morphology of SAE 521000 steel ring after wear test at 400 N load (sliding speed 0.4 m SK’).
Fig. 11. Longitudinal cross-section through the worn surface revealing the extent of plastic deformation during wear of the T&,Ni,,Fe, at 400 N load (sliding speed 0.4 m SC’).
and the steel as a function of load and sliding velocity are summarized in Tables 2 and 3, respectively.
4. Discussion One of the most important results obtained in the present study is that the intermetallic Ti,,Ni,,Fe, shows 2-5% as much wear as steel during dry sliding wear against the SAE 52100 steel although the intermetallic has a lower bulk hardness (300 kg mm-‘) in comparison to that of the SAE 52100 bearing steel (900 kg mm-‘). EDS results (Fig. 8) as well as worn surface metallography (Fig. 7) clearly reveal that a layer containing mainly iron and its oxides forms on the wear surface during the wear process. The presence of iron oxides lowers the wear damage by acting as solid lubricants as well as reducing the metal to metal contact [11,12]. In the beginning of the wear process, Ti,,Ni,,Fe, makes
Fig. 12. Tapered section of the intermetallic block tested under 400 N load (l-2” to the worn surface), depicting (a) finger type morphology below the contact surface and (b) microhardness values obtained at various locations at the surface in (a) (sliding speed 0.4 m SC’).
direct contact with the steel ring and therefore shows high wear rates (Fig. 4(a)). When a layer of iron oxides starts to form on the surface of the Ti,,Ni,,Fe,, the wear rates show a decrease with increasing sliding distance until this layer reaches an equilibrium thickness to give steady state wear rates. The intermetallic shows in general a high oxidation resistance and therefore can be expected to show lower oxidative wear rates than the steel counterface. In ductile materials wear processes generate large plastic strains beneath the contact surfaces. It is shown by cold rolling experiments (Fig. 3) that Ti,,Ni,,Fe, can considerably work harden at high strains. In fact, the microhardness measurements taken on the contact surface, in locations outside the mechanically mixed finger like features, indicate that the surface hardness of T&N&Fe, can reach up to 400 kg mm-‘, an increase of about 30% relative to the bulk hardness. Thus, the hardening of the contact surfaces during wear may increase the wear resistance of Ti,,Ni,,Fe,. It is im-
J. Singh, A.T. A&as
/ Wear 181-183
(1995) 302-311
0
i
2
3
309
4
X-RAY
5
6
ENERGY
7
a
s
B
BlO
8
9
10
(KsV)
I”“““’
j 0
12
3
4
X-RAY
2
3 X-RAY
4
5
6
ENERGY
7
B
a
10
0
1
2
(KeV)
3 X-RAY
5
6
ENERGY
4
5
7
(KeV)
6
ENERGY
7
10
(KsV)
Fig. 13. Transverse cross-section through a mechanicahy mixed layer (finger type morphology) along with EDS analysis results. Locations where the analysis was performed arc indicated by letters a-d. Table 2 Summary of wear characteristics
as a function of load Regime I
Regime II
Regime III
1 x 10-5-1 x 10-d 2x 10-4-2 x lo-’
-
Wear rates (mm’ m-‘)
Ti,$Ji,,Fe3 ME52100
1 x10-6-1 x10-5 2x1o-5-2x1o-4
Worn surface
Tis,,Ni4,Fe3
??
SAE 52100 Wear debris
smooth ?? reddish brown (oxide) layer ?? no plastic deformation ??
smooth and shiny
fine powder reddish brown ?? oxide (FerO3)
shallow groves 0 dark brown (oxide) layer ?? plastic deformation ?? hardened layer ?? fine groves, shiny ??
fine powder dark brown 0 oxide (Fe203,Fe304)
>2x10-3 rough with deep groves black layer (Fe rich) ?? heavy plastic deformation ?? formation of hardened fingers ?? uneven shiny ??
??
black and mixed large metallic Fe and ?? fine oxide (FezOp and Fea04)
??
??
??
??
??
??
portant to note that the cold worked inter-metallic material can retain its strength at elevated temperatures (Table l), so that, once formed, the hard surface layers are not expected to soften while the contact surface temperature increases during the sliding process. High hardness values (Fig. 12(b)) recorded in the mechanically mixed layers may be attributed to a structure which consists of a mixture of iron oxide debris with the base material. Formation of mechanically mixed layers on the worn surfaces have been reported by
several researchers [13,14]. TEM examination by Rainforth et al. [14] showed that the submicroscopic iron oxide particles penetrate up to a depth of several tens of microns and intimately mix with the base material. This is in agreement with the observation that the layers of a mechanical mixture of iron oxides and intermetallic compounds have penetrated into the T&N&Fe, up to a depth of about 80 pm. Although, details of the formation of iron oxide rich finger like features are not clear, their formation can be the result
310
.I. Singh, A.T. Alps
Table 3 Summary
of wear characteristics
as a function
of sliding
Sliding
speed
/ Wear 181-183
speed <2.5
m SC’ (pre-transition)
Wear rates (mm’ m-‘)
TiSONi,,Fe, SAE. 52100
1 x 10-5-2x 3x 1o-4-6x
Worn
Ti,,Ni,,Fe,
?? reddish brown 0 fine groves ?? subsurface deformation ?? hardened layer
Wear
surface
debris
(1995) 302-311
-
speed
>3.0
m SK’ (post-transition)
2x10-4-3x10-4 3x10-5-2x10-5
10-S 1o-4
black patches rough with deep groves ?? minimal deformation ?? subsurface cracks ??
??
0 fine powder reddish brown ?? iron oxides (Fe203,Fe304)
??
??
of the penetration of the mechanically mixed material into the subsurface material. Don et al. [15] have indicated that, whenever the surface hardness H is greater than the hardness h of the base material, i.e. H/h > 1.2, the hardened layer will penetrate and sink into the subsurface region. The wear of T&N&Fe, can be described as mild at low sliding speeds and severe at high sliding speeds (greater than 3.0 m s-l) as indicated by the shape of the wear rate vs. the sliding velocity plot of Fig. 6. At low sliding speeds, the low wear rates of this alloy can be attributed to the work hardening of the surface layers and the formation of a hardened wear resistant layer which consists of iron oxides and the base alloy. However, above a critical sliding speed of about 3.0 m s-‘, the protective tribolayers cease to form. Metallographic examination of the subsurface region showed no sign of formation of finger like features or iron oxide veins at high sliding speeds. Instead, cracks emanating from the bottom of the surface groves running normal to the wear surface were observed. The wear process, under these conditions, produces metallic debris in the 100 pm size range. It appears that the ductility of T&N&Fe, is adversely effected by the high strain rates caused by high sliding speeds. Subsurface crack formation could also be the result of thermal fatigue due to higher temperature gradients and thermal cycling of the surface which is no longer protected by the tribolayers. The possibility of the martensitic transformation has also been considered as an alternative explanation of the transition to severe wear since the martensite of T&N&Fe3 is brittle and much softer [16,17]. XRD analysis performed on the debris generated at 4 m s-l did not reveal martensite. This was nevertheless expected since martensite, if it could be induced during sliding, would be pseudo-elastic and would transform back to the austenite phase once the stress is removed. However, formation of stress induced martensite during sliding is unlikely. Firstly, the MS temperature of T&N&Fe, is - 140 “C [lo], whereas, the temperature near the contact surface (at a sliding speed of 4 m
Sliding
??
large (= 100 wrn) metallic (Ti5aNt4,Fe3)
s-l and 20 N load) was measured to be 170 “C only after 10 min (by inserting a 0.2 mm thick ungrounded thermocouple about 50 pm below the contact surface). Subsurface temperature increases with speed and the stresses needed for the stress induced martensite can easily reach 3-5 GPa [16,18]. Secondly, mechanical twinning rather than martensitic transformation is preferred at ambient and high temperatures [5]. The wear of the SAF 52100 steel undergoes a change in wear mechanism at a sliding speed of 2.6 m s-l. At this sliding speed, the steel appears to undergo a transition from mild oxidative type wear to severe oxidative wear, resulting in a large drop in its wear rates. Such a transition has been related to the formation of an oxide layer on the steel surface itself when the temperature of the steel is sufficiently high to reform oxide at a rate faster than it is removed. Increase in the sliding speed increases the surface temperature and therefore the oxidation rate of iron.
5. Conclusions 1.
2.
3.
A thermomechanical treatment (45% cold work followed by annealing at 450 “C for 10 min) can be used to strengthen T&N&Fe, intermetallic alloy to obtain an optimum combination of the yield strength (830 MPa) and ductility (15% tensile elongation). The strengthened T&N&Fe, alloy shows excellent wear resistance with wear rates between 10m6 and 10e5 mm3 m-’ in the load range 2-400 N, when tested for dry sliding wear against a SAE 52100 steel. These wear rates of the T&N&Fe, are only 2-5% of those of the SAE 52100 bearing steel counterface. During sliding wear, surface layers of the Ti,,Ni,,Fe, work harden and also form finger like features with very high hardness. These layers, containing a mechanical mixture of iron oxides and constituents of the base material, form beneath the contact surface. The surfaces themselves are covered with Fe,O,
J. Singh, A.T. Alpas / Wear 181-183
4.
and Fe,O, oxidized wear products transferred from the counterface. At high sliding speeds, higher than 3.0 m s-l, T&N&Fe, suffers severe wear. However, the steel counterface shows a transition from mild oxidative wear to severe oxidative wear when the sliding speed exceeds 2.6 m s-l.
Acknowledgements The authors gratefully acknowledge financial support provided by Natural Science and Engineering Research Council of Canada (NSERC).
References [l] [2]
S.P. Gupta and A.A. Johnson, Trans. Jpn. Inst. Met., 14 (1973) 292. C.M. Wayman, J. Met, 6 (1980) 129.
(1995) 302-311
311
and R.J. Biermann, J. Met., 2 (1988) 32. [31 SM. Tuominen [41 E. Goo, T. Duerig, K. Melton and R. Sinclair, Acta MetaN., 33 (1985) 1725. [51 W.J. Moberly, J.L. Proft, T.W. Duerig and R. Sinclair, Acta Metall. Mater., 38 (1990) 2601. Fl K.N. Melton and 0. Mercier, Acta Metall., 27 (1979) 137. Titanium Zirconium, 27 (1979) [71 Y. Suzuki and T. Kuroyanagi, 67. Wear, 146 (1991) 219. [81 Y. Shida and Y. Sugimoto, [91 P. Clayton, Wear, 162-164 (1993) 202. [lOI E. Goo and R. Sinclair, Acfa Metall., 33 (1985) 1717. T.F.J. Quinn, Br. L Appl. Phys., 13 (1962) 33. PI P21 H. Kato, T.S. Eyre and B. Ralph, Acta Metall. Mater., 42 (1994) 1703. P31 P. Heilmann, J. Don, T.C. Sun, D.A. Rigney and W.A. Glaeser, Wear, 91 (1983) 171. R. Stevens and J. Nutting, Phi/ox Msg., A66 1141 W.M. Rainforth, (1992) 621. 1151 J. Don, T.C. Sun and D.A. Rigney, Wear, 91 (1983) 191. J. Perkins and J.M. Johnson, Ser. Metall., 9 1161 G.R. Edwards, (1975) 1167. P71 S. Miyazaki, T. Imai, Y. Igo and K. Otsuka, Metall. Trans., 17A (1986) 115. P81 F. Takei, T. Miura, S. Miyazaki, S. Kimura, K. Otsuka and Y. Suzuki, Ser. Metall., 17 (1983) 987.