Journal of Materials Processing Technology 123 (2002) 13±17

Grinding of alumina/aluminum composites

Zhaowei Zhong*, Nguyen Phu Hung
School of Mechanical and Production Engineering, Nanyang Technological University, Nanyang Avenue, Singapore 639798, Singapore
Received 29 October 2000

Abstract

With the anticipated widespread usage of metal matrix composites (MMCs) in the near future, the machinability of high performance
MMCs needs to be understood. This paper reports research results obtained from the grinding of aluminum-based MMCs reinforced with
Al2O3 particles using grinding wheels having SiC in a vitri®ed matrix and diamond in a resin-bonded matrix. The issues discussed are surface
roughness, grinding force, type and size of the abrasives, grinding conditions, and the consequential sub-surface integrity. The study used
grinding speeds of 1100±2200 m/min, a grinding depth of 15 mm for rough grinding and 1 mm for ®ne grinding, and cross-feeds of 3 and 1 mm
for rough and ®ne grinding respectively, while maintaining a constant table feed-rate of 20.8 m/min. The surface integrity of the ground
surfaces and sub-surfaces were analyzed using a scanning electron microscope (SEM) and a pro®lometer. The surface ®nish values, Ra, were
scattered in the range 0.15±0.70 mm for the rough-ground samples, whilst a narrower range of 0.20±0.35 mm was achieved for the ®ne-ground
samples. Smearing of aluminum on the ground surfaces was seen for rough grinding, but was negligible for ®ne grinding because all the Al2O3
particles of the ground surfaces were clearly visible when observed with the SEM. Grinding using a 3000-grit diamond wheel at depth of cut of
1 mm produced many ductile streaks on the Al2O3 particles. Both the Al2O3 particles and aluminum matrix were removed by micro
machining. There were no cracks and defects found on the ground surfaces. There was almost no sub-surface damage, except for a rare
cracked particle being found. Rough grinding with a SiC wheel followed by ®ne grinding with a ®ne-grit diamond wheel is recommended for
the grinding of alumina/aluminum composites. # 2002 Published by Elsevier Science B.V.

Keywords: Metal matrix composites; Grinding; Ductile streaks

1. Introduction A study on the machining of high performance MMCs,

therefore, becomes important, especially where mass pro-
Aluminum-based metal matrix composites (MMCs) rein- duction is involved.
forced with ceramic particles are known as advanced mate- Sensitive cost and fabrication challenges including
rials for their high speci®c strength, high wear resistance, machining must be overcome for successful application
and good damping characteristics. Methods to produce these of these composites. The surface ®nish and surface integrity
composites and studies on their mechanical properties have are important for surface-sensitive parts subjected to fatigue
gained popularity [1]. MMCs are primarily used in the or creep. Machine-induced sub-surface damage could have
astronautic, military and automobile industries. In addition, serious consequences. However, many machining methods
the sporting goods industry has also been in the forefront of often result in cracking, splintering and pulling-out of
MMCs development, capitalizing on the materials' high reinforcement particles. Sub-surface damage resulted from
speci®c properties. There is also a growing interest in other conventional and unconventional processes such as turning,
industries, such as the shipping industry [2]. drilling, milling, electrical discharge machining, abrasive jet
Reports on the machining of aluminum-based MMCs machining, and laser machining [5]. A proper process such
reinforced with ceramic particles [3±7] are still scarce. as grinding to obtain a good surface ®nish and damage-free
Despite many advantages, full implementation of MMCs surfaces is crucial for the application of the materials.
is cost-prohibitive, partially due to the material's poor Ductile-mode machining of brittle materials, such as
machinability. Although near-net shape MMC products Al2O3 and SiC, has resulted in many innovative applications.
can be manufactured, ®nal ®nishing may still be needed It was reported that by having a critical depth of cut and with
for the ®nal designed dimensions and required surface ®nish. ¯attened grains slightly protruding from the surface of the
grinding wheel, ¯awless machining, free of brittle fracture, is
*
Corresponding author. possible [8]. Evidence of plastic ¯ow with Al2O3, Si3N4, and
E-mail address: [email protected] (Z. Zhong). SiC was shown and a model based on the combination of two

0924-0136/02/$ ± see front matter # 2002 Published by Elsevier Science B.V.

PII: S 0 9 2 4 - 0 1 3 6 ( 0 2 ) 0 0 0 7 5 - 4
14 Z. Zhong, N.P. Hung / Journal of Materials Processing Technology 123 (2002) 13±17

theories was proposed [9]. However, reports on the ductile- Table 2

mode machining of aluminum-based MMCs reinforced with Grinding wheels and grinding conditions for rough and fine grinding
ceramic particles are still very scarce. Therefore, further Rough grinding Fine grinding
studies on the ductile-mode machining of the materials are
Grinding wheel
required. Grain Green SiC Diamond
This paper reports research results obtained from the Grit size 80 3000
grinding of aluminum-based MMCs reinforced with Bond Vitrified-bond Resin-bond
Al2O3 particles using grinding wheels having SiC in a Diameter (mm) 350 350
Width (mm) 38 10
vitri®ed matrix and diamond in a resin-bonded matrix.
The issues discussed are surface roughness, grinding force, Dressing stick
type and size of the abrasives, grinding conditions and the Grain Green SiC WA
Grit size 60 320
consequential sub-surface integrity.
Grinding speed
Low speed (m/min) 1100 1100
Medium speed (m/min) 1650 1650
2. Experiments High speed (m/min) 2200 2200
Depth of cut (mm) 15 1
The MMCs chosen for the grinding experiments were Feed-rate (m/min) 20.8 20.8
2618/Al2O3/10p (W2F10A) and 2618/Al2O320p (W2F20A) Cross-feed (mm) 3 1
detailed in Table 1. The grinding wheels used and grinding
conditions are shown in Table 2. Six 2618/Al2O3/10p and six
2618/Al2O3/20p workpieces were milled to the same size. The surface roughness of the ground MMCs in the
Three 2618/Al2O3/10p and three 2618/Al2O3/20p work- grinding (table feed) direction and cross-feed direction
pieces were rough-ground with an 80-grit vitri®ed-bond was measured using a Tokyo Seimitsu roughness tester.
SiC grinding wheel at low, medium and high grinding speed. The cut-off was 0.8 mm and the evaluation length was
The remaining three pieces of 2618/Al2O3/10p and 2618/ 4 mm. The average value was calculated from three mea-
Al2O3/20p workpieces were ®rst rough-ground with the SiC sured values in the same direction on each ground surface.
grinding wheel at low grinding speed, and then ®ne-ground The surface integrity of the ground surfaces was analyzed
with a 3000-grit resin-bonded diamond wheel at low, med- using a Cambridge±Leica scanning electron microscope
ium and high grinding speed. (SEM). Furthermore, the ground surfaces were cleaned with
Grinding experiments were carried out on an Okamoto alcohol, protected with epoxy, then sectioned with a dia-
precision surface-grinding machine (PSG-64DX). A Fuji mond saw on a Buehler Isomet metallographic cutter. The
inverter (FVR-G75) was attached to the machine so that the specimens were cold vacuum mounted, ground, polished,
main spindle of the machine was capable of being changed. and then etched in a solution of 2 ml HF 48%† ‡ 3 ml
A dynamometer (Kistler9257A) was mounted on the table of concentrated HCl ‡ 5 ml concentrated
the grinding machine to measure the grinding forces. The HNO3 ‡ 190 ml H2O. Sub-surface damage was assessed
dynamometer was connected to charge ampli®ers (Kis- by observing the etched samples using an SEM.
tler5011) and the measured grinding forces were recorded
usinga chartrecorder(Yokogawa LR8100).Thegrindingforce
reported here is the force perpendicular to a ground surface. 3. Results and discussion
A SiC wheel mounted on a brake-controlled truing device
and a single diamond dresser were used for truing the grinding The surface roughness values of rough and ®ne-ground
wheels. SiC and WA dressing sticks were used for dressing MMCs are shown in Figs. 1 and 2, respectively. The values
the SiC and diamond grinding wheels, respectively. Dres- of the maximum grinding force measured during the rough
sing was carried out before every grinding experiment. and ®ne grinding experiments are shown in Figs. 3 and

Table 1
MMC workpieces used for the grinding experiments

Matrix 2618 Aluminum alloy

Reinforcement 10 vol.% Al2O3 particulate for 2618/Al2O3/10p (W2F10A); Al2O3 particle size: 9.3 mm
20 vol.% Al2O3 particulate for 2618/Al2O3/20p (W2F20A); Al2O3 particle size: 21 mm
Process Direct extrusion with a flat face die
Extrusion ratio 20:1
Billet temperature 420±430 8C
Extrusion speed 3 m/min
Dimensions 19 mm length†  17 mm width†  17 mm height†
 Z. Zhong, N.P. Hung / Journal of Materials Processing Technology 123 (2002) 13±17 15

Fig. 3. Grinding forces for rough grinding (80-grit vitrified-bond SiC

wheel; depth of cut: 15 mm; cross-feed: 3 mm; feed-rate: 20.8 m/min).

Fig. 1. Roughness values of ground MMC surfaces. Rough grinding

(80-grit vitrified-bond SiC wheel; depth of cut: 15 mm; cross-feed: 3 mm;
feed-rate: 20.8 m/min).

4, respectively. Figs. 5 and 6 show the top and the cross-

section of the rough-ground surfaces, while Figs. 7, 8 and 9
show those of the ®ne-ground surfaces.
The roughness values of the rough-ground MMC surfaces
in the cross-feed direction were higher than those in the
grinding (table feed) direction. However, the effect of the
measurement direction on the surface roughness values of
the ®ne-ground MMC surfaces was not signi®cant. The
surface ®nish values, Ra, were scattered in the range of
0.15±0.70 mm for the rough-ground samples, whilst a nar-
rower range of 0.20±0.35 mm was achieved for ®ne-ground Fig. 4. Grinding forces for fine grinding (3000-grit resin-bond diamond
samples. The surfaces ground by the 80-grit SiC wheel at wheel; depth of cut: 1 mm; cross-feed: 1 mm; feed-rate: 20.8 m/min).
speed of 1100 and 1650 m/min at a depth of cut of 15 mm
had roughness values close to those of surfaces produced by
the 3000-grit diamond grinding wheel at a depth of cut of
1 mm: this was due to the smearing of the aluminum matrix.
Smearing of aluminum on the ground surfaces was seen for

Fig. 2. Roughness values of ground MMC surfaces. Fine grinding Fig. 5. An SEM micrograph of a ground MMC surface (2618/Al2O3/20p).
(3000-grit resin-bond diamond wheel; depth of cut: 1 mm; cross-feed: Rough grinding (80-grit vitrified-bond SiC wheel; grinding speed: 2200 m/
1 mm; feed-rate: 20.8 m/min). min; depth of cut: 15 mm, cross-feed: 3 mm; feed-rate: 20.8 m/min).
16 Z. Zhong, N.P. Hung / Journal of Materials Processing Technology 123 (2002) 13±17

Fig. 8. Sub-surface of ground MMC (2618/Al2O3/20p). Fine grinding

(3000-grit resin-bond diamond wheel; grinding speed: 1100 m/min; depth
of cut: 1 mm; cross-feed: 1 mm; feed-rate: 20.8 m/min).

Fig. 6. Sub-surface of ground MMC (2618/Al2O3/20p). The arrows show

commonly found cracks Al2O3 particles. Rough grinding (80-grit vitrified- speed and reduced the grinding force component perpendi-
bond SiC wheel; grinding speed: 2200 m/min; depth of cut: 15 mm; cross- cular to the ground surface. In the case of ®ne grinding,
feed: 3 mm; feed-rate: 20.8 m/min). because the depth of cut was 1 mm, the thermal effect might
be negligible. Further investigation is needed to better
understand the micro machining mechanism of both the
rough grinding, but was negligible for ®ne grinding because soft matrix and hard, brittle particles at the same time.
all the Al2O3 particles of the ground surfaces were clearly However, from Figs. 3 and 4 it can be seen clearly that
visible when observed with the SEM. the grinding force required for grinding the MMCs with
The maximum grinding force decreased with increasing 20 vol.% Al2O3 particles is always higher than that for
grinding speed for the rough grinding, but increased with grinding the MMCs with 10 vol.% Al2O3 particles.
increasing grinding speed for the ®ne grinding. This could be Figs. 5 and 6 show that cracks of the Al2O3 particles occur
due to the different abrasives, grit sizes and depths of cut on and under the rough-ground surfaces. As mentioned
used for the rough- and ®ne-grinding experiments, the above, the surfaces rough-ground at speed of 1100 and
thermally induced softened matrix at high speed for rough 1650 m/min have roughness values close to those of ®ne-
grinding, etc. For example, because the depth of cut was ground surfaces. However, SEM pictures show that the
15 mm for the rough grinding of Al2O3 particles (particle surface topographies of the rough- and ®ne-ground surfaces
size: 9.3 or 21 mm), more heat was generated in the defor- with close surface roughness values are signi®cantly differ-
mation zone. This softened the matrix at the higher grinding ent. No Al2O3 particles were seen on the rough-ground
surfaces, except for some small holes showing fractured
Al2O3 particles, as seen in Fig. 5. Almost the whole of
the surfaces were smeared with the soft aluminum matrix.

Fig. 9. The sub-surface of ground MMC (2618/Al2O3/20p). The arrow

Fig. 7. SEM micrograph of ground MMC surface (2618/Al2O3/20p). Fine shows the very rare crack of an Al2O3 particle. Fine grinding (3000-grit
grinding (3000-grit resin-bond diamond wheel; grinding speed: 1100 m/ resin-bond diamond wheel; grinding speed: 2200 m/min; depth of cut:
min; depth of cut: 1 mm; cross-feed: 1 mm; feed-rate: 20.8 m/min). 1 mm; cross-feed: 1 mm; feed-rate: 20.8 m/min).
 Z. Zhong, N.P. Hung / Journal of Materials Processing Technology 123 (2002) 13±17 17

Some of the aluminum chips were back-transferred on to the ground surfaces were clearly visible when observed with
top of the surfaces. an SEM.
SiC wheels are much cheaper than diamond wheels, the Grinding using the 3,000-grit diamond wheel at depth of
cost ratio being roughly 1:10±20. Because the depth of cut cut of 1 mm produced many ductile streaks on the Al2O3
and cross-feed used were 15 and 3 times those for ®ne particles. Both the Al2O3 particles and aluminum matrix
grinding respectively, the stock removal was substantial and were removed by micro machining. There were no cracks
the grinding time was much shorter as compared to that for and defects found on the ground surfaces. There was almost
®ne grinding. Hence, the potential of using SiC wheels at no sub-surface damage except for a rare cracked particle
least for rough grinding is high. Rough grinding parameters being found.
and dressing frequency should be optimized to make rough The potential of using SiC wheels at least for the rough
grinding using SiC wheels more attractive. grinding of alumina/aluminum composites is high, because
As shown in Fig. 7, grinding of the alumina/aluminum SiC grains are harder than Al2O3 reinforcing particles and
composite 2618/Al2O3/20p (20 vol.% Al2O3) using the ®ne- much less expensive than diamond grains. Rough grinding
grit diamond wheel at 1 mm in-feed (depth of grinding) with a SiC wheel followed by ®ne grinding with a ®ne-grit
produced visible ductile streaks on the Al2O3 particles. Both diamond wheel is recommended for the grinding of alumina/
the matrix and the Al2O3 particles were removed by micro aluminum composites.
machining because the ductile grinding marks were clearly
seen on the Al2O3 particles. There were no cracks and
defects found on the ground surfaces. There was almost References
no sub-surface damage as shown in Fig. 8, except for a very
rare cracked particle, as shown in Fig. 9. [1] M.J. Tan, L.H. Koh, K.A. Khor, F.Y.C. Boey, Y. Murakoshi, T. Sano, J.
Mater. Process. Technol. 37 (1993) 391±403.
[2] N.L. Loh, in: Proceedings of the International Symposium on
High Performance Metal Matrix Composites, Japan, 1994, pp. 11±
4. Summary of results 12.
[3] A.R. Chambers, S.E. Stephens, J. Mater. Sci. Eng. A 135 (1991) 287.
Grinding experiments using both 80-grit SiC in a vitri®ed [4] L.A. Looney, J.M. Monaghan, P. O'Reilly, D.M.R. Taplin, J. Mater.
Process. Technol. 33 (1992) 453.
matrix and 3000-grit diamond in a resin-bonded matrix were
[5] N.P. Hung, F.Y.C. Boey, K.A. Phua, H.F. Lee, J. Mater. Process.
performed on aluminum-based MMCs reinforced with Technol. 56 (1996) 966±977.
Al2O3 particles. [6] N.P. Hung, Z.W. Zhong, C.H. Zhong, in: Proceedings of the Fourth
The surface ®nish values, Ra, were scattered in the range Conference on Composites Engineering, Hawaii, 1997, pp. 459±460.
of 0.15±0.70 mm for the rough-ground samples, whilst a [7] N.P. Hung, Z.W. Zhong, C.H. Zhong, J. Mater. Manuf. Process. 12 (6)
(1997) 1075±1091.
narrower range of 0.20±0.35 mm was achieved for ®ne-
[8] W. Konig, V. Sinhoff, SPIE, Lens and Optical Systems Design, 1992,
ground samples. Smearing of aluminum on the ground pp. 778±788.
surfaces was seen in rough grinding, but was negligible [9] K. Kitajima, G.Q. Cai, N. Kumagai, Y. Tanaka, H.W. Zheng, Ann.
for ®ne grinding because all the Al2O3 particles of the CIRP 41 (1) (1992) 367±371.