MO OVERLAYS

NATIONAL TRIBOLOGY CONFERENCE 24-26 September 2003

THE ANNALS OF UNIVERSITY “DUNĂREA DE JOS“ OF GALAŢI FASCICLE VIII, TRIBOLOGY 2003 ISSN 1221-4590

93

TRIBOLOGICAL PROPERTIES OF AG/GRAPHITE/MO OVERLAYS FOR PLAIN BEARINGS PREPARED BY ECR-DC SPUTTERING Cristian Lungu1,2,Kunihiko Iwasaki2 1

National Institute for Lasers, Plasma and Radiation Physics, Bucharest, Romania, 2 Japan Ultra-high Temperature Materials Research Institute, Ube-city, Japan [email protected]

ABSTRACT In order to improve the electron cyclotron – direct current (ECR-DC) sputtering process, Ag/Graphite/Mo overlays were produced by using composite target made by sintering and compared with those prepared by the ”ring method”. Wear and frictional testing was carried out with a CSEM (Switzerland) ball-on-disk tribometer. The counter material was SUJ2 ball bearings with 6 mm in diameter. Measurements were made in air (45-48% relative humidity) at room temperature (25-29 oC) and both in dry and in lubricated (SAE#30 oil) conditions. The properties of the prepared overlays were studied and compared with those of the Ag/Graphite overlays. Optical microscopic observation of the wear scars of overlays after testing at different sliding speeds in dry and lubricated conditions shows that the smooth wear scar appears at a low sliding speed. At a sliding speed of 0.1 ms-1 a plastic flow can be observed along the sliding direction. KEYWORDS: Ag, graphite, ECR-DC sputtering, plain bearing, dry and lubricated sliding.

1. INTRODUCTION

potential applied to target; 800 V, DC current; 280 – 500 mA.

Ag-matrix coatings acting as anti-friction layers for the engine plain bearings were found to have a superior combination of mechanical and physical properties such as high toughness, high fatigue resistance, low coefficient of friction in dry and lubricated sliding and excellent resistance against wear and seizure [1-2]. They are also ecologically friendly, replacing Pb base overlays currently in use. Addition of graphite and metal dopants to Ag matrix overlays was found to reduce the coefficient of friction in dry and lubricated sliding, respectively [2].

2. EXPERIMENTAL METHODS 2.1 Overlay deposition Ag matrix overlays were prepared using an ECR-DC hybrid system presented in detail elsewhere [3]. The Pb-free bronze substrates (30 mm x 30 mm x 2 mm) were ground with 600 grit emery paper to have a rough uniform initial surface. The deposition time was 7.2 ks and the processing conditions were: sputtering distances; 20; 30 and 40 mm, pressure; 0,093 Pa, Ar flow rate; 160 mm3s-1, microwave power; 400 W, microwave frequency; 2,45 GHz,

Fig. 1 The principle of the ring assembling. The target consists of a stack of Mo, Gr and Ag rings as shown in figure 1. The thickness and the internal and external diameters of each ring were 5 mm, 80 mm and 90 mm, respectively. This method is very useful for the preliminary experiments to

94


determine the optimum compositions and is called “ring method” here. The ratios of the number of Mo, Gr and Ag rings were set at 0/2/8; 1/2/7; 2/2/6 and 3/2/5. The number of Gr rings was kept at 2 for all the cases here because the ratio Gr/Ag = 2/8 had been found to be the best in the previous work [5]. An electrically grounded flange made of stainless steel was set at the edges of the target as shown in the lower part of figure 1 to prevent the plasma from squeezing into the backside of the target. This gives rise to a shadowing effect on the rings located at the edges of the target to decrease their effective areas. Then the ratio of the effective areas depends not only on that of the number of the rings but also on theirs stacking positions or sequence. By making use of this shadowing effect, however, it is possible to prepare an overlay with low Mo concentration as will be shown later. A composite target was made by sintering 97 mass% Ag + 2 mass% graphite + 1 mass% Mo powders of 10-50µm mean diameter. The powders were milled for 72ks and compacted by pressing them at 60MPa in a special mold of 60 mm in diameter. The compacts were cold-isostatic-pressed (CIP) at 294MPa for 1.8ks. Finnaly the compound was sintered in air at 873K for 900s. The sintered composites were shaped in a cylindrical mold with a 45 mm radius in order to be fitted in the target holder. The thickness of the overlays prepared by using a pile of rings and a composite target was in the range of 12 – 17µm, commonly used thickness for the overlay of a three-layer engine bearing [4]. Wear and frictional testing was carried out with a CSEM (Switzerland) ball-on-disk tribometer. The counter material was SUJ2 ball bearings with 6 mm in diameter. Measurements were made in air (45-48% relative humidity) at room temperature (25-29 oC) and both in dry and in lubricated conditions using a SAE#30-oil. Load of 5N and sliding radii between 2 and 5 mm were chosen. The sliding speed was 0.01 ms-1 and 0.1 ms-1 in dry condition and 0.1 ms-1 in lubricated condition. Wear rate was calculated from the cross-sectional area of the wear scar measured with a profilometer. The calculated wear volume was divided then by load and sliding distance. Micro Vickers hardness was measured using an Akashi microhardness tester. Loads of 9.8 mN, 98 mN and 15 s indentation time were used. After indentation, the indentation imprints were analyzed optically. A thrust-type test was carried out using a specifically designed specimen. The flat-to-flat sliding was performed keeping in Hertzian contact a coronal surface of the substrate and a shaft counter part. During the test, the applied load was increased stepwise by 3MPa at 300s time interval. The specimen temperature and the coefficient of friction were continuously monitored. A sharp raise of the coefficient of friction or specimen temperature was noticed


3. RESULTS AND DISCUSSION 3.1 Overlays prepared by ring method 3.1.1 Composition of the overlays Chemical composition of the overlays prepared by the ring method was found to depend not only on the number of the constituent target rings but also on theirs stacking sequence. The rings located near the center of the cylindrical target is more effective than those located near the edges because the former are exposed to the plasma more effectively than the latter due to the presence of the grounded flange as was described before (Fig. 1). Four kinds of overlays (a), (b), (c) and (d) were prepared here by using the stacking number and sequence shown in the second column of Table 1, where 1/2/7, for example, indicates 1 Mo, 2 Gr and 7 Ag rings were stacked together from left to right as shown in figure 1 The concentrations of Mo, C and Ag are summarized in the table. It is clearly seen that Mo concentration increased more rapidly than that predicted from the increase in the number of Mo rings. The concentrations of C and Ag were kept almost constant except the case (d). The sudden change observed in the concentrations of Mo, C and Ag in the case of (d) comes not only from the shadowing effect but also from the staking position of the Mo rings that was close to the substrate compared to that of the Ag rings. A trace of Mo detected in the overlay (a) may be due to the contamination of the former experiments.

(a) (b) (c) (d)

Table 1. Composition of the overlays (a) – (d). Mo C Ag Mo/Gr/Ag mass% mass mass % % 0/2/8 0.06 1.7 98.2 1/2/7 0.8 1.6 97.6 2/2/6 4.2 1.5 94.3 3/2/5 45.0 1.0 54.0

3.1.2 Mechanical and tribological propetries The results of the micro-hardness measurements are summarized in figure 2, where the average values of more than 3 measurements are plotted with error bars. It is to be noted here that the overlay (b) with Mo has lower hardness than the overlay (a) without Mo. This is quite unexpected because Mo is known to have higher hardness than Ag. This is explained by the formation of a significant porous structure by the addition of Mo as shown in figure 3(b), where the indentation imprints after the microhardness test of the overlays (a) – (d) are presented, though its formation mechanism has not been known yet. This porous structure is the main cause for the softening. The porosity, however, becomes smaller when the Mo concentration is


increased further as in (c) and (d), and microhardness drastically increases. 400 300 250 200 150

95

overlay surface [4] is conside-red to dominate the friction behavior at this stage. After the initial stage is over, the energy necessary to plow the extremity of the cone-structured surface as well as to overcome the abrasive wear leads to an increase in the coefficient of friction [3, 13]. This corresponds to the cases of (c) and (d) with high Mo contents. The overlay (b) exhibits the most stable and the lowest coefficient of friction both at low (0.01ms-1) and at high (0.1ms-1) sliding speeds.

100 50 0

(a)

(b)

(c)

(d)

0/2/8

1/2/7

2/2/6

3/2/5

Mo/Gr/Ag rings

Fig. 2 Micro-Vickers hardness of the deposited films using Mo/Gr/Ag ring combinations: (a); 0/2/8, (b); 1/2/7, (c); 2/2/6, (d); 3/2/5. (a)

(b)

(c)

(d)

(d)

Coefficient of friction

Vickers hardness, HV

350


(a)

0,6

(a)

(d) (c)

(c)

0,4

(b)

(b) 0,2

-1

Speed: 0.01 ms Radius: 30 mm 0

500

1000

-1

Speed: 0.1 ms Radius: 40 mm 1500

2000

Sliding time, s 30 µm

Fig. 3 Indentation imprints of the developed overlays after the indentation test (Load: 98mN). The brittleness of the film (c) and (d) also increases, as can be deduced from the cracks formed around the indentation imprints. The softening that was observed in the overlay (b) is considered to be very beneficial for the improvement of conformability and embeddability of the plain bearing overlay. It is to be noted here that the overlay (b) with Mo has lower hardness than the overlay (a) without Mo. This is quite unexpected because Mo is known to have higher hardness than Ag. This is explained by the formation of a significant porous structure by the addition of Mo as shown in figure 3b though its formation mechanism has not been known yet. This porous structure is the main cause for the softening. The porosity, however, becomes smaller when the Mo concentration is increased further as in (c) and (d), and microhardness drastically increases. The brittleness of the film (c) and (d) also increases, as can be deduced from the cracks formed around the indentation imprints. The softening that was observed in the overlay (b) is considered to be very beneficial for the improvement of conformability and embeddability of the plain bearing overlay. Figure 4 shows the typical results of the ball-ondisk tests for all kinds of overlays in dry sliding at 0.01 ms-1 and 0.1 ms-1. It is clearly seen that the results depend strongly on the Mo concentration. The coefficient of friction in dry sliding decreased with the increase in the Mo concentration at the initial runningin stage. The low adhesive force between the hard ball and the top of the relatively soft cone-structured

Fig. 4 Coefficient of friction behavior during ball-on disk test of the films prepared using Mo/Gr/Ag ring combinations: (a); 0/2/8, (b); 1/2/7, (c); 2/2/6, (d); 3/2/5. Temperature: 24oC, relative humidity of air: 45%, ball diameter: 6 mm, applied load: 5N.

3.2 Overlays prepared by composite targets 3.2.1

Target preparation

For the industrial use of the ECR-DC sputtering method is necessary to have composite sputtering targets in order to insure the reproducibility of the process. For this purpose we prepared Mo-Gr-Ag composite targets, used them for sputtering and made overlays characterization. The flow chart of composite target preparation is shown in figure 5. Mo-Gr-Ag powders with 20-50 µm grain size were milled during 20 h in dry air. The chosen mixtures were: A: 97mass%Ag, 2mass%Gr, 1mass%Mo and B: 96mass%Ag, 2mass%Gr, 2mass%Mo. After milling, the mixtures were compacted using molds having 60 mm in diameter. After 60MPa pressing, 30-g powder was compacted as about 2 mm thickness disks. The disks were pressed using the CIP (cold isostatic pressing) method at 294MPa, 30 minutes. After that the compacts were sintered and cylindrical shaped in dry air at about 600OC for 15 minutes. Figure 6 shows such a process when a cylindrical load is pressed on the compact disc settled in a cylindrical mold made by Ti.


96

THE COMPOSITE TARGET METHOD Powders

Ag

Graphite


Figure 6 shows the photograph of the developed composite target mounted into the ECR-DC water cooled target holder after a sputtering process.

Mo

3.2.2 Milling: 20 h Pressing: 60 MPa

Cold Isostatic Pressing (294 MPa) Sintering (600oC, 15 min)

Ball-on-disk tribological tests

Ag-matrix including Mo and Gr were deposited on rectangular substrates (30 mm x 30 mm x 2mm size) made of rough (#600 grit) bronze as well as on the special designed substrates for the thrusttype seizure resistance tests. Deposition time was 120 minutes and the processing conditions were: sputtering distance: 2-4cm, pressure: 0.092Pa (7 x 104 Torr), Ar flow rate: 10 sccm, microwave power: 400 W, DC bias: 800 V, DC current: 350 - 400mA.

Fig .5 Flow chart of composite target preparation

Fig. 7 Wear scar after tribological test of the overlays prepared using 97mass%Ag, 2mass%Gr and 1mass%Mo powder Fig. 5 Cylindrical shaping of the compact disc: 90 mm radius of curvature

An optical micrograph of the wear scars of overlays prepared using 97mass%Ag, 2mass%Gr, 1mass%Mo powder combination, 30 mm sputtering distance after testing at different sliding speeds in dry and lubricated conditions are shown in figure 7.


0,8 0,6 40 mm 30 mm 20 mm Bronze Dry sliding Sliding speed: 0.1 m/s

0,4 0,2 0,0 0

200

400

600

800

1000

Sliding time, s

Fig. 6 Composite target appearance after the sputtering process. The prepared disks were cut in cylindrical sectors about 45mm in length and then assembled into the target holder of the ECR-DC sputtering system.

Fig. 8 Frictional characteristics in dry sliding, high sliding speed, of the coatings prepared at 20, 30 and 40 mm sputtering distance, high speed. At a low sliding speed the wear tracks are smooth. At a sliding speed of 10 mms-1 a plastic flow along the sliding direction can be observed. In lubricated conditions at a speed of 100mms-1 the



0,8 0,6

-1

80

-15

2

Wear rate, x 10 m N

surface was smooth too, with small grooves along the sliding direction. The coefficients of friction of the overlays prepared at 20, 30 and 40mm sputtering distance were in the range of 0.65 ± 0.05, lower than that of the bronze substrate (0.80 ± 0.05) as can be seen in figure 8 for the high sliding speed test (0.1ms-1) and in figure 9 for higher sliding sped test (0.01ms-1) performed in dry condition. In lubricated conditions, the coefficients of friction of the coatings were lower and more stable that that of bronze substrate, after a certain sliding distance (50m), as can be seen in figure 10. The coefficient of friction of bronze was low at the beginning due to the sliding on the top of the “hills” present on the surface. Latter, a close contact with material increases the coefficient of friction.


97

-1

0.01 ms (dry) -1 0.1 ms (dry) -1 Lubricated (0.1ms ) Load: 5 N, 45% Relative humidity

60 40 3 2 1 0

Bronze

20 mm

30 mm

40 mm

Sputtering distance

Fig. 11 Wear rates in dry and lubricated sliding, high sliding speed and low sliding speed of the coatings prepared at 20, 30 and 40 mm sputtering distance. As comparison, the wear rates relative to the bronze substrate are shown, too.

3.2.3 Thrust-type tribological tests 40 mm 30 mm 20 mm Bronze Dry sliding Sliding speed: 0.01 m/s

0,4 0,2 0,0 0

200

400

600

800

Sliding time, s

1000 1200

Fig. 9 Frictional characteristics in dry sliding of the coatings. Sliding at low speed.

0.110

In the thrust type test, the sliding contact occurs only on the segmented circular part with 22 and 27.2mm in internal and in external diameters, respectively. The schematic diagram of the test equipment is shown in figure 12. The test conditions are listed in the caption of the same figure. Typical temperature and torque monitoring chart during the test is shown in figure 13, together with those ones of the bronze substrate. The results of the test shown in figure 14 reveal the higher value of seizure resistance of the composite overlay deposited at 40mm sputtering distance, in connection with the lower roughness of the overlay.

(a)


Bronze

0.105

(b)

0.100 0.095

(a): sputtering distance = 20 mm (b): sputtering distance = 30 mm (c): sputtering distance = 40 mm -1 Sliding speed: 0.1 ms Load: 5 N, Lubricant: Oil SAE #30

(c)

0.090 200

400 600 Sliding time, s

800

1000

Fig. 10 Frictional characteristics in lubricated sliding, high sliding speed, of the coatings prepared at 20, 30 and 40 mm sputtering distance. Figure 11 shows the effect of the sliding speed and sliding conditions on the wear rates. The wear rates and the scar surface appearance reveal a mild wear both in dry and in lubricated conditions of the overlays prepared using the composite target method.

Fig. 11 Schematic diagram of the fundamental seizure test. Test conditions: Shaft roughness: Rmax: 1 µm, Bearing size: ∅27.2x ∅ 22mm, Sliding velocity: 2m/s (1500 rpm), Oil grade: SAE #30, Oil supply rate: 20 ml/min, Load: increased with 3MPa at every 30 minutes.


98


Figure 14 shows the results of the thrust-type tests of special designed specimens coated with the Mo-Gr-Ag overlays using composite targets. In this figure the results of two kinds of used composite target are shown. Analysed by X-ray fluorescence (XRF), the Mo content in the prepared overlay was found to 0.8mass% in one case and 1.8mass% in the second one. From the obtained results we can infer the beneficial influence of Mo addition in small quantities (about 0.8mass%) to the Ag-Gr overlay. As found by using the “ring method”, in dry sliding Mo can led to the formation of a solid lubricant as MoOx and in lubricated condition MoS2 is expected to be formed by a tribochemical reaction.

4. CONCLUSIONS

Seizure resistance, MPa

Fig. 12 Specimen temperature and the torque evolution during the thrust type test of the overlay prepared using the composite target. A comparison with the bronze substrate is shown. 16 14 12 10 8 6 4 2

0 Sputt. dist., mm: 20

30

40

Rmax, µ:

0.72

0.60

0.79

Fig. 13 Thrust-type tests results for the composite overlay using 97mass%Ag, 1mass%Mo and 2mass%Gr, deposited at 20 mm, 30 mm and 40mm.

Seizure resistance, MPa

27 24 21

REFERENCES

18 15 12 9 6 3 0

Ag-matrix overlays containing Mo as dopant were successfully prepared by ECR plasma sputtering enhanced by DC hollow cathode discharge, using Ag, graphite and Mo as sputtering target. The Ag/Graphite/Mo overlays were found to have low and stable coefficient of friction tested in lubricated conditions. Mo addition to the Ag/graphite overlay decreased the hardness of the film (at low Mo concentration as 0.8 mass%) and then increased the hardness. The coefficient of friction and the wear rate of Mo/Gr/Ag overlays were reduced from 0.6 ± 0.5 to 0.3 ± 0.5 and from 20 x 10-15 m2N-1 to 12 x 10-15 m2N1 respectively compared with that of 2/8 Gr/Ag overlays. The best combination of the Mo, Gr and Ag rings was determined to be 1/2/7, which corresponds to a 0.8 mass% Mo concentration into the film. The films prepared using sintered composite targets (Ag + 2mass%Gr+1mass%Mo and Ag + 2mass%Gr+2mass%Mo) showed excellent tribological properties. Films that included 0.8mass%Mo exhibited maximum seizure resistance, lowest coefficient of friction and wear rate.

Ag

0.8Mo-Gr-Ag

1.8Mo-Gr-Ag

Coating composition

Fig. 14 Seizure resistance of the (0.8Mo-Gr-Ag) and (1.8Mo-Gr-Ag) composite overlays, compared to that of the pure Ag overlay.

1. Lungu C. P., Iwasaki K., Kawachi T., Ishikawa H., 2001, Proc. 2nd World Tribology Congress, Vienna, Sep., pp 497-503. 2. Lungu C. P., IwasakiK., 2002, Proc. JAST Tribology Conference, Tokyo, May, pp 285-286. 3. Lungu C. P., IwasakiK. , Vacuum, 66 (2002) 385-391 4. Ishikawa H., Nomura K., Mizuno Y., Michioka H., Y. Fuwa Y., Yasuhara S., 1996, SAE paper 960988, 1. 5. Lungu C. P., Iwasaki K., Kawachi T., Ishikawa H., 2001, in F. Franek, W. J. Bartz, A. Pauschitz (eds.) Tribology, The Austrian Tribology Society, Vienna, pp. 291. 6. Iwasaki K., Lungu C. P., Takayanagi S., Ohkawa K., 2002, “Influence of graphite addition on the tribological properties of plain bearings overlays”, Int. J. of Applied Mechanics and Engineering, vol. 7, pp. 351-356