JOURNAL

OF MATERIALS

SCIENCE

LETTERS

13 ( 1 9 9 4 ) 6 6 3 - 6 6 7

Wear studies of fibre-reinforced ceramic matrix composites S. M. B L E A Y , V. D. S C O T T , B. H A R R I S

School of Materials Science, University of Bath, Claverton Down, Bath, BA2 7A Y, UK

Glass and glass-ceramic systems reinforced with ceramic fibres are possible replacements [1-4] for metallic materials in gas turbines because of their good specific strength and stiffness, high-temperature stability and corrosion resistance but, surprisingly for materials envisaged for use in an engine with moving parts, only limited wear studies have hitherto been carried out. Some data are available on Pyrex glass :reinforced with carbon fibres [5, 6], showing that the material exhibited the good wear resistance of glass while retaining the lubricating capability of tl~e carbon fibre, that its friction coefficient depended upon the modulus of the carbon fibres, higher-modulus fibres giving lower values of friction coefficient, and that fibre orientation affected wear rate. Unfortunately, however, composites rein:~orced with carbon fibres have restricted use because of oxidation of the fibres and, for high-temperature applications, other ceramicbased composite materials have been considered, such as alumina and silicon nitride with whiskers of silicon carbide [7, 8] and silicon nitride containing particles of titanium carbide, nitride or boride or boron nitride [9]. In this Ietter, 'we first report on the wear behaviour, as a function of load (10 N to 50 N), of two fibre-reinforced ceramic-matrix composites (CMC) and then we examine the effect on wear of raising the test temperature from ambient to 600 °C. The two composite materials used in this study consisted of calcium magnesium aluminosilicate (CMA6; Ceramic Developments Midlands Ltd, UK) reinforced with Tyranno (Ube Ltd, Japan) fibres (designation CMAS/Tyr) and calcium aluminosilicate reinforced with Nicalon (Nippon Carbon Ltd, Japan) fibres (CAS//Nic); they were both crossply material with a nominal fibre volume fraction of 0.4. Wear tests were carried out in a reciprocating pin-on-plate machine fitted with a furnace. Wear pins, 6 mm x 11 mm x 3.5 mm, were cut from a plate of (0, 90)4s CMAS/Tyr, while the baseplates, 35mmx20mmx2nnn, were made of either (0,90)3~ CAS/Nic or unreinforced CAS. The surfaces were prepared by grinding on successively finer abrasive down to 6/zm grade. The first series of tests was performed at room temperature with a baseplate of CAS/Nic and applied loads of 10 N, 30 N and 50 N. The second set of tests was carried out on the two types of baseplate, with a load of 30 N and test temperatures of 300 °C and 600 °C as well as ambient. In all cases the sliding speed was 25 m m s -1 and the duration of the test was 2 h. Wear was monitored by weighing the pin and baseplate before 0261-8028 © 1994 Chapman & Hall

and after testing and calculating wear rate from sample dimensions. Surfaces of the wear tracks on pin and baseplate were examined by scanning electron microscopy (SEM), and wear debris was collected for analysis by SEM and energy-dispersive X-ray spectrometry (EDS). Sections through the wear pin were also prepared for study. Data from room-temperature tests, shown in Fig. 1, show that the wear rates of pin and baseplate increased with increase in load, with the accompaniment of a corresponding increase in experimental scatter as illustrated by the error bars. Comparison of the pin before (Fig. 2) and after wear testing at a load of 10 N (Fig. 3) revealed matrix fragmentation, which was possibly associated with sub-surface fibres. The corresponding wear track in the CAS/Nic baseplate showed removal of matrix asperities to 0.1.2

T

0.15

0.10 0.08

E

/V 4.-" 4," /l/

,~ 0.06 o

/1/

g O.lOg i11 o

¢.

i£ 0.04

T

T ~

10

0.05

// // // // // //

0.02 0

// //

0 30

50

Load (N)

Figure1 Effectof load on the wearof CMAS/Tyrpin (unshaded) against CAS/Nicbaseplate (shaded) at room temperature; error bars indicatespreadof results.

Figure2 SEM micrograph, showing surface of as-received CMAS/Tyrpin.

663

Figure 3 SEM micrograph, showing wear surface of CMASfFyr pin, 10 N load.

reveal fibres (Fig. 4). Damage of the pin surface was more extensive with a 30 N load, with crack deflection by fibres and fragmentation of fibres (Fig. 5). Although the width of wear track on the baseplate was greater and matrix cracking more severe, the operative wear mechanism was still the removal of matrix asperities and the exposure of underlying fibres. There was clear evidence of fibre removal, a polished section through the wear track showing residual "sockets" left behind in the matrix. The wear rate was noticeably faster when fibres were aligned parallel to the direction of sliding than when they were end-on, as illustrated in sections through

wear pins tested under loads of 10 N, 30 N and 50 N (Fig. 6). Damage to the baseplate was also more extensive at the higher load, with a deeper wear track and removal of the original surface to expose fresh fibre and matrix. The development of subsurface damage could be seen in sections through the wear track. Wear debris were identified by EDS as agglomerates of CMAS and fine angular particles of Tyranno fibre, both types of debris thus originating from the pin (Fig. 7). In tests at 300 °C, wear rates of pin and baseplate (CAS/Nic) were significantly higher (Fig. 8). The surface of the pin revealed breakage of longitudinal fibres, with debris trapped in end-on fibres which then became "smeared" over the surface (Fig. 9). Differences in wear rate between adjacent plies were again evident, although the surface was more

Figure 4 SEM micrograph, showing wear surface of CAS/Nic baseplate, 10 N load.

Figure 5 SEM micrograph, showing wear surface of CMAS/Tyr pin, 30 N load.

664

Figure6 Optical micrograph, showing polished sections of CMAS/Tyr pins tested under (a) 10 N, (b) 30 N and (c) 50 N loads.

Figure 7 SEM micrograph, showing wear debris from 30 N wear

test.

Figure lO SEM micrograph, showing wear debris, 300 °C test,

30 N load. 0.6

1.5

0.5 A

E 1.0

~" 0.4 E

e~ "o

0.3

o t-

-o.5

~-. 0.2

0.1 0 I 30

300 600 Test t e m p e r a t u r e ("C)

Figure8 Effect of temperature on wear of CMAS/Tyr pins (unshaded) against CAS/Nic baseplate (shaded), 30 N load.

Figure 11 SEM micrograph, showing section through CMAS/Tyr pin showing matrix cracking, 600 °C test, 30 N load.

Pins tested against a baseplate of unreinforced CAS suffered substantially less wear than those run against CAS/Nic composite (see Fig. 12), although the surface appeared somewhat similar. W e a r on the CAS baseplate was much the same as for the CAS/Nic baseplate although the appearance of the wear track differed with temperature, with largescale matrix cracks at r o o m t e m p e r a t u r e (Fig. 13), a

0.10 -

--

7~

0.08 I

Figure 9 SEM micrographs, showing wear surface of CMAS/Tyr

pin, 300 °C test, 30 N load.

E E 0.06 -

rounded than that produced in the roomt e m p e r a t u r e test. The wear track in the baseplate extended into the lower plies, the fine debris at the b o t t o m of the groove consisting of material from the baseplate and fragments of Nicalon fibre (Fig. 10). W e a r of the plate increased when the test was conducted at 600 °C, but wear of the pin decreased as the surface b e c a m e smeared with debris. A polished section shows matrix cracking around fibres and oxidation of exposed fibres (Fig. 11).

E Q. (I1

0.5

I ¢-

1.0

// /f // // // // // // // // // // //

I

// //

0.04 -

-a o

//

0.02 -

Z Z Z

0 20

300

600

Test temperature (°C) Figure 12 Effect of temperature on wear of CMAS/Tyr pins (unshaded) against unreinforced CAS baseplate (shaded), 30 N load. 665

Figure 13 SEM micrograph, showing wear surface of C A S baseplate after r o o m temperature test.

more granular structure and some smearing at 300 °C (Fig. 14) and a fine granular appearance at 600 °C (Fig. 15). These-results clearly show that the increased severity of wear which accompanied an increase in load was associated mainly with matrix fragmentation, lift-out of longitudinal fibres and chipping of end-on fibres. While in some regions fibres could be seen to deflect cracks, it was evident that at some interfaces debonding occurred so as to initiate cracks and give rise to a fatigue wear situation. The faster wear observed when fibres were aligned parallel to the sliding direction, compared with end-on plies, was more noticeable at the higher loads, implying

that fibre breakage and lift-out occurred at a faster rate than chipping of fibre ends. Raising the wear test temperature significantly increased the wear rates of pin and baseplate, with smearing of debris across the pin and the development of a more rounded surface. The influence of temperature on wear may be related to the role played by the carbon-rich layer known to be present at the fibre-matrix interface of these composites [10, 11]. During roonl-temperature wear, interfaces are exposed by matrix fragmentation and, even though the carbon-rich layer is thin, sufficient carbon is available to assist lubrication. At higher temperatures, the carbon becomes oxidized to form gaseous products and its lubricating effect disappears. As a consequence, wear debris builds up between the pin and baseplate, leading to threebody abrasive wear and a substantially higher wear rate compared with the fatigue-induced wear at low temperatures. Furthermore, since debris can no longer escape from grooves in the baseplate, it is ground progressively finer and smeared across the sliding surfaces. It therefore follows that the worsening wear at 600 °C is due to the faster oxidation rate of carbon. Wear on the pin (CMAS/Tyr) was less when run against a baseplate of unreinforced CAS compared with a CAS/Nic baseplate. This was to be expected since the pins were significantly harder than unreinforced CAS material. The influence of temperature on pin wear was much the same whether CAS and CAS/Nic baseplates were used, although the carbonrich interfaces assisting lubrication were, in the former case, contributed solely by the CMAS/Tyr pin. However, the temperature versus wear characteristics of the two baseplate materials were different and, at high temperatures, granular pluck-out occurred in the CAS baseplate, whereas large-scale fragmentation, initiated by cracks at fibre-matrix interfaces, remained the major damage mechanism in the CAS/Nicalon baseplate.

Acknowledgements Figure 14 S E M micrograph, showing wear surface of CAS baseplate after 300 °C test.

The authors thank the SERC for support, Berkeley Nuclear Laboratories for the wear rig, and Mr J. J. R. Davies of AEA, Harwell, for providing the CMAS/Tyranno material.

References 1. 2. 3. 4. 5. 6.

Figure 15 S E M micrograph, showing wear surface of C A S baseplate after 600 °C test.

666

7.

J . J . B R E N N A N and K. M. P R E W O , J. Mater. Sci. 17 (1982) 2371. K . M . P R E W O , ibid. 22 (1987) 2695. K . M . P R E W O , B. J O H N S O N and S. S T A R R E T T , ibid. 24 (1989) 1373. J. F. M A N D E L L , D. H. G R A N D E and J. J A C O B S , J. Engng Gas Turb. Power, Trans A S M E 109 (1987) 267. E. M I N F O R D and K. M. P R E W O , Wear 102 (1985) 253. z . L U , K. F R E I D R I C H , W. P A N N H O R S T and J. HEINZ, J. Mater. Sci. Lett. 12 (1993) 173. C . S . Y U S T , J. M. L E I T N A K E R and C. E. D E V O R E , Wear 122 (1988) 151,

8.

H. L I U , M. E. F I N E , H. S. C H E N G and A. L. G E I G E R ,

J. Amer. Ceram. Soc. 76 (1993) 325. 9. 10.

A. SKOPP, M. W O Y D T and K. H. H A B I G , Tribol. Int. 25 (1992) 61. R . F . C O O P E R and K. C H Y U N G , J. Mater. Sci. 22 (1987) 3148.

11.

S. M. B L E A Y , V, D. S C O T T , B. H A R R I S , R. G. C O O K E and F. A. H A B I B , J. Mater. Sci. 27 (1992) 2811.

R e c e i v e d 19 O c t o b e r and accepted 1 N o v e m b e r 1993

667