Heinrich J.G., Aldinger F. (Eds.) Ceramic Materials and Components for Engines

This Page Intentionally Left Blank

DESIGN

OF

WEAR RESISTANT POLYCRYSTALLINE ALUMINA

D.

Galusek", F.L. Riley**, R. Brydson**

(*)Institute

of

Inorganic Chemistry, Slovak Academy

of

Sciences, M. Razusa 10,911 05

Trench, Slovak Republic

(**)

Department

of

Materials, School

of

Process, Environmental, and Materials Engineering,

University

of

Leeds, LS2 9

JT

Leeds, UK

ABSTRACT

Alumina is often used in applications requiring

high wear resistance. Ultra fine ultra pure aluminas

(with mean grain sizes

c

1 pm) exhibit the highest

wear resistance, wearing primarily tribochemically.

Coarse-grained alumina with the mean grain

size

>

1 pm wears predominantly by grain boundary

microfracture, with rates an order of magnitude faster

than submicrometre ones. Carefully designed

microstructure and tailored grain boundaries, however,

allow significant improvement of the wear resistance of

coarse-grained aluminas. The paper attempts to identify

the parameters responsible for improved resistance of

liquid phase sintered alumina with grain size

>

1

pm

against the wet erosive wear.

INTRODUCTION

Polycrystalline alumina ceramics are widely used

for industrial applications where high resistance to

wear is required. Ultra fine ultra pure aluminas (with

mean grain sizes

<

1 pm) exhibit the highest wear

resistance. They wear very slowly, primarily by

tribochemical processes. However, fabrication of

aluminas with mean grain sizes below one micron

requires expensive high purity starting powders,

carefully designed processing routes, or special

preparation procedures (e.g. sol-gel), which are usually

not suitable for industrial production due to high

production and raw materials' costs.

The grain size of the most standard commercial

aluminas is usually well over 1 pm. Such materials are

prepared by liquid phase sintering, and normally

contain several percent of a combination of calcium

and magnesium silicates.

'

The wear rates for wide

range of wear modes (including erosive and abrasive

wear, cutting, and grinding) of such materials are

known to increase significantly with the grain size.

-6

The main wear mechanism is microfracture and crack

interlinking, leading to grain detachment and the

development of rough surfaces. In the case of pure

alumina the fracture mode appears to be predominantly

intergranular, suggesting fracture follows preferen-

tially the grain boundaries.

The use of silicate based sintering additives leads

to significant improvements in the wear resistance of

alumina, and this is well recognized industri-

ally. Recent laboratory studies using model high purity

aluminas of controlled composition have shown the

generally beneficial effects of Group I1 metal silicate

additions on the wear resistance of alumina

materials. The magnesium silicate appears to confer the

greatest benefit.

'

We report here results of work on magnesium

silicate (MgSi03) densified alumina, with the focus on

fully dense polycrystalline materials containing up to

10

%

by weight of magnesium silicate of 1:l

stoichiometry. The effects of additive quantity, mean

alumina grain size, and residual stresses have been

examined in detail.

EXPERIMENTAL

Alumina powders (CL3000SG, Alcoa International

Ltd., Worcester, UK) with or without sintering

additives were hot-pressed

(HP)

in a graphite die at 20

MPa and at various temperatures and times depending

on the amount of additives to obtain fully dense discs

(6 mm thick and of 25 mm diameter). The hot-pressing

conditions

are

listed in Table 1. Magnesium silicate

(MS) was added by mixing the alumina powder with

isopropanol solutions of magnesium nitrate

Mg(N03)2.4H20 (AnalaR grade, BDH Ltd., Poole,

UK) and tetraethylorthosilicate (TEOS, AnalaR grade,

BDH Ltd., Poole, UK). The amounts of these materials

were calculated to achieve final compositions with 0.5,

1,

5,

or

10

%

by weight of MgSi03 with MgO

:

SiOz

molar ratio of 1

:

1. An aqueous solution of ammonium

hydroxide (10 wt

%)

was then added to precipitate

magnesium hydroxide, and to hydrolyse the TEOS and

the suspension was dried under an infrared lamp. The

dry powder was calcined

30

min at

900

"C.

Table

1

Hot-pressing conditions and mean grain

sizes (MGS) of the materials studied.

MS

HP

conditions

MGS

wt.

%

T,OC

t,min

pm

A1

0

1620 45 2.5

A2

0

1620

70

3.4

M0.5 0.5 1500

70

2.1

M1 1 1500 60 2.0

M5

5

1450

30

1.9

M10 10 1450 20 1.8

To

induce grain growth, selected hot-pressed

materials were subsequently annealed without pressure

in the hot-pressing die at temperatures 50 "C higher

than the temperature of hot-pressing for

3

or 6 h, or for

6

h

at 1620 "C.

Surface layers of the hot-pressed discs were

removed by silicon carbide paper (240 mesh powder)

and bulk densities

(p)

were determined by mercury

163

immersion. Fragments of hot-pressed discs were milled

in a tungsten carbide puck mill and used for a

XRD

measurement using a powder diffi-actometer (Philips

APD1700) operating at 40 kV with Cub radiation

within the range of angles

5"

to

70".

To avoid possible effects of porosity on wear rate,

only materials with bulk density

>

98

%

of theoretical

were used for further studies. Discs were weighed to

k

0.1

mg and worn under carefully standardised

conditions in an abrasive slurry consisting

of

750

g of

0.5

to 1

mm

fused alumina grit (Morganite 954 grog,

Morganite Ceramics Ltd, Neston, UK) in 250 cm3

distilled water. The disc was clamped horizontally in a

specially designed cylindrical polyurethane block

so

as

to expose approximately one half of the disc

surface. The block was attached to the shaft of an

attritor mill, and rotated at

500

r.p.m., giving an

effective alumina disc perimeter velocity of

1.9 ms-'. The final disc weight was measured to

rf:

0.1 mg, and the weight losses occurring between the

standard times

of

2 h and

6

h

were used to calculate a

wear rate. Full details of the equipment and

experimental procedures are given elsewhere.

After wear rate measurements, discs were cut by a

diamond wheel and polished by diamond paste to

1

pm

finish. Polished specimens were etched thermally (pure

alumina at 1500

"C

for

30

min), or chemically (MgSiO,

sintered alumina, boiling phosphoric acid for

30

s),

to

reveal grain boundaries and examined by

SEM. Micrographs were analysed by standard image

analysis software (KS

400,

Kontron Electronics

GmbH, Germany) to give mean grain sizes and grain

size distributions.

For TEM examination specimens were core

drilled and sliced to prepare 3

mm

diameter discs of

500

pm thickness. The discs were further ground,

polished, dimpled and ion-milled (Gatan

PIPS

low

angle ion beam thinner) until transparent to the electron

beam, and coated with carbon. Specimens were

examined using a transmission electron microscope

(Philips CM20

200

keV) fitted with a scanning

(STEM) unit, ultra thin window EDX detector (Oxford

Instruments) and PEELS (Gatan).

Cr3+ luminescence spectra were determined on

polished surfaces using a Raman optical microprobe

(Renishaw Model

2000)

and incident radiation from a

HeNe laser at

633

nm

(15797.79 cm-') in order to

assess thermal expansion-anisotropy stresses in

polycrystalline aluminas. The incident light was

focused using a

5x

objective giving a spot size much

greater than the alumina grain size and a polished

sapphire single crystal was used as the unstrained

reference. Spectra were taken from at least

6

different

positions on each specimen, including the sapphire

single crystal, which was measured before and after

each alumina specimen. The details of data analysis are

given elsewhere.

8

50

-

'q

40-

E

f

30-

K

r

3

20-

10-

RESULTS

Hot pressing followed by high temperature

annealing allowed

to

prepare pure and magnesium

silicate-sintered aluminas with a range of grain sizes.

For all materials the wear rates increased with mean

alumina grain size, and confirmed the strong influence

of alumina grain size on wear rate. The measured

values for the pure alumina agreed satisfactorily with

those of the earlier study. The incorporation of small

amounts of MgSi03 markedly reduced the wear rate.

Figure

1

shows wear rate as a function

of

mean

alumina grain size for different levels of MgSi03

addition.

....

P

0

pureAI,O,[6

-.-pure

AI,O,

MgSiO,:

-V-

0.5

wt.%

-0-1

Wt.%

-0-

5

Wt.%

-A-

1

O

Wt.%

01

.

I

-

I

.

I

.

I

.

0

2

4

6

8

10

De

1

Ctm

Fig.

1

The wear rate of pure and magnesium silicate

sintered aluminas vs. the mean grain size.

The SEM examination revealed strikingly

different morphologies of the wear surfaces of

MS-doped specimens, compared to the pure aluminas.

The facetting was an apparent feature

of

the wear

surface of the pure alumina, and indicated clean grain

detachment resulting from intergranular fracture, grain

boundary cleavage, and crack interlinking. On the other

hand, the roughness

of

the wear surfaces of magnesium

silicate sintered materials decreases progressively with

increasing magnesium silicate content. The smoother

areas in the magnesium silicate sintered materials

appear to be the combined result of smooth

transgranular cleavage coupled with subsequent

tribochemical polishing, in the absence of significant

intergranular fracture leading to the detachment

of

entire grains. (Fig 2)

TEM examination of materials containing

MgSi03 showed a

-

2 nm thick continuous amorphous

film at all alumina-alumina boundaries examined.

Excess glass seggregated in triple grain junctions

(triple pockets).

A

typical feature is shown in Figure

3.

The composition of the silicate varied between the two-

grain boundary zone and the larger triple point regions,

as had earlier been found for CaSi03 doped alumina

materials.

lo

The two-grain boundaries (grain faces)

had roughly equal proportions (expressed as atom

%)

of

Mg and Si, whereas the triple point regions were

164

Fig.

2

SEM micrographs of wear surfaces of polycrystalline aluminas with various contents of magnesium silicate:

O%-a,l%-b,5%-c,lO%-d

rich in

Si

(Si

:

Mg ratio approximately

3

:

1). Alumina

grains in materials containing MgSi03 showed

evidence of considerable stress, in the form of

dislocation networks.

Supplementary

XRD

examinations showed small

amounts of crystalline secondary phases formed by the

reactions of MgSi03 and the alumina matrix. The main

secondary phase in materials hot-pressed for shorter

times were spinel (MgA1204) and sapphirine

(Mg4AlloSizOZ3). High temperature annealing promoted

the formation of mullite (A16Si2013), accompanied by

the disappearance of sapphirine. These secondary

crystalline phases detected by

XRD,

together with a

low thermal expansion silicate film

on

the alumina

grain boundaries, would be expected to lead to residual

stresses on cooling from formation temperature,

because of mismatches

of

thermal expansion

coefficient.

”

Analysis of the Cr3+ photoluminescence spectra

showed obvious differences between the

specimens. The measurements confirmed the presence

of high local fluctuating compressive and tensile

stresses on the level of approximately

450

MPa in the

MgSi03 sintered materials, which were significantly

higher than those in the pure alumina

(-

200

MPa).

DISCUSSION

This work reveals several features of the influence

of the magnesium silicate sintering aid

on

the wear

behaviour of polycrystalline alumina. An addition

of

-1

%

MgSi03 reduced the wear rate more than

50

%.

At the same time the predominant fracture path

switches from inter- to transgranular. Associated with

these aspects is a doubling in the maximum in the

fluctuating residual stresses in the alumina

microstructure. These features are examined

individually.

It

appears that the erosion resistance of the

material is primarily determined by the strength

of

grain boundary. The magnesium alumino-silicate

grain boundary phase clearly has a strong influence

on

the microfracture behaviour of alumina. The greater

difficulty in forming microcracks in the silicate

containing material, together with suppression

of

intergranular fracture, suggests that the silicate has a

marked grain boundary strengthening action. The

calcium silicate densified alumina examined earlier

lo

showed amorphous silicate films at all the two-grain

boundaries,

of

the order of

2

nm thickness. The TEM

examinations of the magnesium silicate densified

material used here also showed continuous films of

amorphous silicate at two-grain boundaries. The

excess over the amount needed to form the equilibrium

thickness

l3

grain face film will form pockets at three-

grain edges (“triple points”) and grain corners,

interconnected by the thin grain face films. Although

165

no calculations have been reported for MgSi03 sintered

alumina, theoretical modelling of CaSi03-containing

alumina material suggests that a silicate containing

grain boundary of certain compositions should be

stronger than a pure alumina-alumina boundary.

This is essentially because of the ability of the cations

and their associated oxygen atoms to assist the

relaxation of grain interface stresses caused by

misalignment of Si-0 and Al-0 bonds.

The analysis of the Cr3' piezoluminiscence spectra

suggests that the high fluctuating local stresses in the

magnesium silicate densified aluminas are not caused

by thermal contraction anisotropy of the alumina grains

but by thermal expansion mismatch between the

alumina and localised pockets of intergranular

silicate.

A

silica rich glass will have a much lower

thermal expansion coefficient than the alumina.

It has earlier been assumed that hoop stresses

across grain boundaries generated by thermal

expansion mismatch are important factors influencing

crack pro agation during the wear of two-phase

materials.

'

Large glass pockets, or particles of an

appropriate intergranular phase, located at boundaries

can be expected to generate such stresses. In the

material, assuming a relatively rigid glass matrix, the

alumina grains would be expected to develop internal

tensile radial and hoop stresses during cooling from

sintering temperatures

(-

1600

"C).

In practice, the

glass matrix is likely to have a much lower modulus

than the alumina and some stress relaxation will occur,

particularly at temperatures above T, by viscous

flow. However, as the Cr3+ photoluminescence spectra

show, the fluctuating residual stresses are still

considerable, and much larger than those in pure

polycrystalline alumina arising from thermal expansion

anisotropy. The maximum stresses result from the

isolated pockets of silicate at grain edges and

corners. Moreover, because the silicate is the

continuous phase, hoop compressive stresses (and

radial tensile stresses) along the alumina grain

boundaries will be developed. These provide the basis

for a second grain boundary strengthening action of the

magnesium silicate.

Despite the fact that the addition of magnesium

silicate alters the fracture mode from inter- to

intragranular, the influence of alumina grain size on the

wear rate is preserved. The explanation is based on an

older model of Davidge and Riley developed for

high purity aluminas where fracture occurs primarily

intergranularly. For transgranular fracture it is

assumed that the mean overall crack velocity is given

by the mean grain size and by the time required for the

crack to cross an average grain and the two-grain

boundary. If we assume that the propagation of the

crack through the grain is very fast, then the overall

velocity of crack propagation is primarily controlled by

the time the crack needs for crossing the boundary. In

other words, the crack propagates slower if it

encounters the grain boundary more often, i.e. if the

alumina grains are finer.

14

15

Fig.

3

TEM micrograph of

5

wt.%

MgSi03-sintered

alumina.

1

-

triple pocket,

2

-

array of dislocations with

a strain fringe towards the triple pocket region,

3

-

grain boundary with continuous layer of magnesium

silicate.

The last question is the influence of the amount of

magnesium silicate on wear. It is obvious, from Fig.

1,

that increasing amount of silicate in the material

increases the wear resistance of polycrystalline

alumina. If one assumes the equilibrium thickness of

the grain boundary film,

l3

then the excess glass

seggregates in triple pockets.

The radial and hoop stresses within a pocket

arising from thermal expansion mismatch between the

alumina and a glass pocket, are independent of the size

of the pocket. However, for larger pockets, the hoop

stress will be maintained at larger distances in the

two-grain boundaries, and the mean hoop stress over

the whole boundary should be larger. Crack

progression occurs only when the material in the

vicinity of the crack tip acquires sufficient strain

energy to overcome local constraints, supplied by a

particularly severe local impact of the eroding particle.

The assumption that the energy required to drive a

crack across a boundary should be related

to

the

magnesium silicate (and thus intergranular phase)

content thus has justification in principle.

The large tensile stresses in the alumina grains of

the magnesium silicate densified material would also

provide a basis for an enhanced rate of tribochemical

wear.

l9

Although no clear evidence for this was

obtained in the present work, the increased smoothness

of wear surface of samples with increasing magnesium

silicate content suggests that this mechanism actually

takes place. It seems, therefore, that the magnesium

silicate intergranular phase influences the wear rate by

two separate mechanisms, both involving the thermal

expansion mismatch stresses. By applying hoop

compressive stresses to the boundaries, and tensile

166

stresses to the alumina grains, it inhibits intergranular

crack propagation and favours transgranular

propagation. By raising the chemical potential

of

the

alumina grains it can also potentially enhance the

underlying rate

of

tribochemical wear.

CONCLUSIONS

The mass loss caused by the wear of

polycrystalline alumina with the mean grain size larger

than approximately 1 pm occurs primarily by

microcrack initiation and propagation, and crack

interlinking followed by grain pull out. The magnesium

silicate intergranular phase influences the wear rate by

a mechanism involving the thermal expansion

mismatch stresses. By applying hoop compressive

stresses to the boundaries, and tensile stresses to the

alumina grains, it alters the fracture mode from

intergranular to transgranular, thus increasing the strain

energy required for the crack propagation. Increasing

magnesium silicate content improves the wear

resistance via the increased average hoop compressive

stressed that strengthen the grain boundaries. Despite

the fact that the microcracks propagate primarily

transgranularly, the grain size dependence of wear is

retained. As a result, the magnesium silicate sintered

alumina with the wear resistance two to three times

higher than that of the pure alumina with a comparable

mean grain size has been prepared.

The work provides certain implications for design

of wear resistant ceramics. Together with fine grain

size, the secondary phases with suitable thermal

expansion that results in formation

of

compressive

hoop stresses across the two grain boundaries in the

course of cooling from the processing temperature

appear to improve largely the wear resistance. Increase

of tribochemical component of wear on expense of

microcracking is beneficial. The Group I1 element

silicate grain boundary phases, continuously

distributing at all two-grain boundaries and forming

triple pockets with low-thermal expansion glass appear

to meet the above requirements.

ACKNOWLEDGEMENT

The work was supported by the NATORoyal Society

Fellowship scheme, by EPSRC Grant GWJ86513, by

the Slovak National Grant Agency under the grant

NOVEGA

2/7055/20 and by the NATO Science for

Peace Programme under the contract No

SfP

974122.

Alcoa International Ltd. donated the alumina powders.

REFERENCES

'R. Morrell, Handbook of Properties of Technical

&

Engineering Ceramics: Part 2 Data Reviews. Section

I: High-alumina Ceramics, R.Morrel1,

HMSO

(London 1987).

*R. W. Rice, B.

K.

Speronello, Effect of microstructure

on rate of machining of ceramics,

J.Am.Ceram.Soc.,

59,330-333 (1976).

3S.

M. Wiederhorn, B.

J.

Hockey, Effect of material

parameters on the erosion resistance

of

brittle

materials,

J.Muter.Sci.,

18, 766-780 (1983).

4

D. B. Marshall,

B.

R. Lawn, R. F. Cook,

Microstructural effects on grinding of alumina and

glass-ceramics,

J.Am. Ceram.Soc.,

70, C139-C140

(1987).

5M. G. Gee,

E.

A. Almond, Effect of surface finish on

the sliding wear of alumina,

J.Mater.Sci.,

25, 296-

310 (1990).

%.

Miranda-Martinez,

R.

W. Davidge, F. L. Riley,

Grain size effects on the wet erosive wear

of

high-

pwity polycrystalline alumina,

Wear,

172, 41-48

(1994).

7

P.

C. Twigg, R. W. Davidge,

S.

G. Roberts, F.

L.

Riley, Factors affecting the wet erosive wear of liquid

phase sintered aluminas, in

Euro ceramics

V, Eds.

J.

Baxter et. al.,

Key

Engineering Materials

Volumes

132-136 Part 3, 1524-1527, Trans Tech Publications

Ltd., Switzerland (1997).

D. Galusek, P.C. Twigg, F.L. Riley, Wet erosion of

liquid phase sintered alumina,

Wear,

233-235,

D. Galusek, R. Brydson, P.C. Twigg, F.L. Riley, A.

Atkinson, Y.-H. Zhang, Wet erosive wear of alumina

densified with magnesium silicate additions,

J.Am. Ceram.Soc..

submitted for Dublication

588-95 (1999)

lo

R. Brydson, P. C: Twigg,

S.

C. khen, F. L. Riley,

X.

Pan, M. Riihle, Microstructure and chemistry of

intergranular glassy films in liquid phase sintered

alumina,

J.Am. Ceram.Soc.,

81,369-379 (1998).

C.A. Powell-Dogan, A.H. Heuer, Microstructure of a

96

%

alumina ceramic: 11, crystallization of high-

magnesia boundary

glasses,

J.Am. Ceram. Soc.,

73,

C.A. Powell-Dogan, A.H. Heuer, M.J. Ready,

K.

Merriam Residual stress-induced-grain pullout in

a 96

%

alumina ceramic,

J.Am.Ceram.Soc.,

74, 646-

49 (1991).

D.R. Clarke, On the equilibrium thickness of

intergranular glass phases in ceramic materials,

J.Am.Ceram.Soc.,

70, 15-22 (1987).

I4S.

Blonski,

S.H.

Garofalini, Atomistic structure of

calcium silicate intergranular films in alumina studied

by molecular dynamics simulation,

J.Am. Cerum.Soc.,

"S. English, W.E.S. Turner, Relationship between

chemical composition and thermal expansion of

glasses,

J.Am.Cerum.Soc.,

10,551-570 (1929).

6p.C.

Twigg,

A.

Castro, R.W. Davidge, F.L. Riley,

S.G. Roberts, Indentation induced crack interaction in

alumina ceramics,

Phil.

Mag.

A,

74, 1245-1252

(1996).

"R.W. Davidge, F.L. Riley, The grain size dependence

of the wear of alumina,

Wear,

186-187,45-49 (1995).

"R.W. Davidge,

T.J.

Green, The strength of two-phase

ceramic materials,

J.Mater.Sci.,

3, 629-634 (1968).

'%.

C. Twigg, R. W. Davidge and F. L. Riley, "The

effects of silicon carbide nano-phase on the

wet

erosive wear of polycrystalline alumina",

J.Eur.Ceram.Soc.,

16,799-802 (1996).

11

3677-83 (1990).

12

13

80,

1997-2004 (1997).

1

167

This Page Intentionally Left Blank

CRACK GROWTH

OF

CERAMIC MATERIALS IN SLIDING CONTACT

Hatsuiko Usami*, Junji Sugishita and Hiroshi Murase

Meijo University,

1-501 Shiogamaguchi, Tempaku, Nagoya, 468-8502, Japan

Key Words: Brittle materials, Friction and wear, Crack growth

ABSTRACT

In order to evaluate effects of crack growth on fric-

tion and wear of brittle materials such as ceramics, static

contact for the crack initiation and flexural strength

measurements were connected to the friction and wear

experiment. Materials used in the present study were

soda lime glass and alumina.

A

testing apparatus having

a reciprocal movement was used for the friction experi-

ment at a constant frequency with various normal loads

in

air. The results showed that the wear amount of the

glass depended on the crack growth near the wear scar.

Critical condition for the crack initiation during the

sliding contact was estimated from the static contact test

and was confirmed that the tangential traction resulted in

the increase in the maximum stress at the trailing edge.

The results of the strength measurement test revealed

that the crack growth rate depended on the testing condi-

tion.

the damage development of the subsurface layer. The

crack initiation is one of the most important factor for

brittle materials such

as

the ceramics because of unsta-

ble failure.

The present study is focused on the effect of the

crack initiation on the wear properties of ceramics dur-

ing sliding friction. Soda lime glass bars were used as

model material. Static contact test to determine the criti-

cal condition for the crack initiation and flexural

strength measurement to evaluate the damage develop-

ment

of

subsurface layer near the wear scar were carried

out. The critical condition for the crack initiation in

sliding contact and the effect of the crack growth on the

wear behaviour are discussed.

2.

EXPERIMENTAL

2.1. Materials

1.

INTRODUCTION

Ceramics are the candidate materials for anti

wear systems because of their high hardness and chemi-

cal stability (1,2). Friction and wear behaviour

of

the

ceramics has been studied past decade years (3).

As

results, basic wear mechanism of the ceramics are simi-

lar to those

of

metals and it is well recognized that the

wear amount is considerably smaller. However, it has

been confirmed that the wear amount becomes greater at

the specific condition. For example, toughness of the

material has often influenced on the wear amount when

the abrasive wear is dominant on the wear mechanism

(4).

Although most of the results due to the sliding wear

of

ceramics are carried out at a low normal load to avoid

the

crack growth,

it

is important to evaluate the fiiction

and wear in the condition where the applied load is

enough for the crack initiation.

Damage development due to friction test is usually

characterized by a wear scar of the surface and the wear

properties of the material is evaluated by the wear amount

of the surface.

As

shown in the crack growth on the wear

track of ball bearings, it is also well known that subsurface

layer below the wear track has damaged

(5).

Therefore, it is

important for evaluating the wear properties to investigate

not only measuring the wear amount but also characterize

Soda lime glass bars and alumina spheres were

mainly used for specimens. Table 1 shows mechanical

properties

of

the specimens. The glass bar had rectangu-

lar shape (3x4~40

mm)

in conformity with JIS

R

1601

(6).

The radius

of

the alumina sphere was

9.5

mm.

Table

1:

Mechanical properties

of

Specimens

Young’s

Glass

Surfaces

of

the specimen were finished by polishing.

The comer of the glass bar was bevelled to eliminate the

stress concentration in strength measurement. Crack

initiation under static contact and flexural strength

measurements were carried out after the friction experi-

ment using the same specimen.

2.2. Testing apparatus

A

testing apparatus having a reciprocal

169

movement was used for the friction experiment. Figure

1

shows a schematic of the testing apparatus. The

alumina sphere was fixed into the holder, then was

attached to the octagonal shape dynamo-meter

consisted from

two

curved beams and

8

strain gages.

The glass bar was fixed to the guide way driven by a

DC

motor with a connecting link.

The

normal load was

applied by mounting dead weights on the upper end of

the dynamo-meter. The experiments were carried out

in

air. The glass bar was made in contact with the

alumina sphere first, and the normal load was applied

then the glass bar was driven. The testing condition

was given

in

Table

2.

The test was interrupted at the

specific number

of

cycles to measure the weight

loss

of

the specimens and to observe the surface.profile

A commercial grade oil was supplied to reduce

the tangential traction at the interface. The oil was

supplied just before the test and did not add during the

test. Since the oil was consumed with the friction, it

was suggested that the contact condition at the interface

had transformed with the increase

in

the number of

cycles.

Applied Stroke Maximum

load amplitude

speed

15,40,10ON 2

mm

2.1

mm/s

Octagonal-shape

dvamo-meter

Number of

cycles

1000

Normal

'

load

//-

Fig.

1

Schematic

of

testing apparatus

2.3. Contact failure

Static contact tests were applied to evaluate the

critical stress

for

the crack initiation

of

the

glass

surface

using an electric-mechanical testing apparatus (AG-IOB,

Shimazu.

Co.

Ltd.). The glass bar was mounted on the

silicon nitride plate. The alumina sphere was put on

the glass bar. The alumina sphere was pushed by the

piston with a constant cross head speed of

0.01

mm/min.

An acoustic emission (AE) sensor was set on the

silicon nitride plate near the

glass

bar. The AE signal

was measured continuously during the test. The

critical load was determined from the AE signal, then

the load was removed. After the experiment, the testing

surface of the glass bars was observed by an optical

microscopy to measure the radius of the ring crack

growth.

2.4. Inert strength measurement

Measurement of fracture strength of the bar

specimen was carried out

after

the friction experiment

in

conformity with

4

point bending test to evaluate

crack growth below the wear scar. Schematic of

testing configuration was shown

in

Fig.

2.

The wear

scar located to the lower surface and was subjected to

the tensile stress. The cross head speed was

0.01

mm/min.

CHS=O.Ol

mm/m

in

I

Wear

scar

Fig. 2 Schematic of flexural strength measurement

3.

RESULTS

3.1. Coefficient of friction

Friction force was measured continuously

during the test and was evaluated as the coefficient of

friction. The coefficient of friction as a function of

number of cycles was shown in Fig.

3.

In

dry conditions,

The coefficient of friction was greater at initial stage

and indicated the maximum value at approximately

50

cycles, then decreased and reached to the steady state

after

200

cycles. The values at maximum and at the

steady state were

1

.O

and

0.6

to

0.7,

respectively.

1

0.8

3

.g

0.6

c

0

.-

-

L4

$0.2

V

iioe4

I

oll

"

'I'

I'

"

'I1

"

I

0

200

400

600

800

1000

Number of Cycles,

n

Fig.

3

Variations of coefficient

of

friction

(Filled points indicate the results in lubricated

condition.

)

170

In

lubricated condition, the coefficient of friction

was approximately

0.1

and was almost constant during

test. The influence of the applied normal load on the

coefficient of friction was small in dry and lubricated

conditions.

3.2.

Wear behavior

The wear amount of the specimen was measured

from the weight

loss

of the specimen and was indicated

as

the value per unit normal load. It was difficult to

evaluate the wear amount of the alumina spheres and

the glass bars tested

in

lubricated condition since the

weight change was too small to measure.

?

m1

E

b

-3

CI

*

r

I

E"

a

L

I

u

25N 40N(L)

3

+40N +l00N

0

200 400 600

800

1000

Number

of

Cycles,

n

Fig.

4

Wear amount as a function

of

number

of

Cycles (Filled points indicate the results

lubricated condition.

)

in

Figure

4

shows the wear amount of the glass bar

as

a function of the number of cycles.

Slopes

in

the

Figure correspond to the specific wear rate. The specific

wear rate was greater at initial stage and reduced with

the increase

in

the number of cycles. After

200

cycles,

it seemed that the specific wear rate reached to the

steady state region. The average values at the initial

stage less than

50

cycles and at the steady state region

were approximately

1

O-*

mm2/N and

1

O-'ommZ/N,

respectively. The transition of the wear rate from the

initial stage to the steady state region occurred at

approximately

50

cycles at

15

N

and

200

cycles at

100

N.

There was small difference

in

the specific wear

rate against the variations of the normal load. As

results, it is estimated that the wear

loss

per unit normal

load after

1000

cycles depends on the number of the

cycles of the transition occurrence.

3.3.

Wear scar and wear debris

Optical micrographs of the surface after the test

in

dry condition were shown

in

Fig.

5.

Many cracks

grew near the wear scar were clearly seen. The cracks

grew on the slant direction against the sliding motion in

the initial stage (Fig. 4a). Since the increase of the

crack growth region was smaller than that of the area of

the wear scar, the ratio of the crack growth region and

wear scar decreased with the increase

in

the number of

cycles (Fig. 4b).

The distance of adjacent cracks was

about

100

pm. Such kinds of crack initiation was

observed all cases

in

dry condition.

In

the lubricated condition (Fig.

6),

cracks were

not observed the initial stage at 40

N

(Fig. 6a), but

initiated after

1000

cycles. At

100

N,

Ring cracks was

seen after

1000

cycles (fig. 6b). The angle of the crack

was different that of obtained

in

the dry condition.

Therefore, it

is

estimated that not only applied normal

loads but also tangential traction corresponding to the

coefficient of friction have influenced on the critical

value for the crack initiation.

(a),

IOON,

n=3

(b),

IOON,

n=

1000

Fig.5 Optical micrographs

of

the surface tested

in

dry

condition

Wear debris was collected from the surface near

the wear scar of glass bars tested

in

dry condition (Fig.

7).

It was difficult to collect the debris

in

lubricated

condition

because the oil was spread and consumed

during the test. The shape of the wear debris was

different depend on the number of cycles. At initial

stage, the shape of the debris was flake whose size was

approximately

100

pm.

Comparing the interval

of

the

cracks on the surface shown

in

Fig.

5,

it seemed that the

large wear debris resulted

in

the interaction of the crack

growth.

At steady state region, the wear debris was

much smaller. Considering the wear behavior, the wear

amount depends on the size of the wear debris.

171

Heinrich J.G., Aldinger F. (Eds.) Ceramic Materials and Components for Engines

Подождите немного. Документ загружается.