Chapter I

INTRODUCTION

Powder metallurgical technique has been adopted by man

to shape different metals. Strangely enough, the history of

powder metallurgy dates back to thousands of years, when many

metallic articles (precious metal articles made by the Incas

and later the Iron Pillar of Delhi) were believed to be powder

r
metallurgy products.

pu
ar tThis art was somehow lost but was again revived only as
h
la t e as 1826 when platinum ingots were compacted from platinum

powder. The major breakthrough was made in the early 1 9 0 0 ’ s when

ag
Kh ig
Coolidge and others made ingots of tungsten powder, which were

made into the filaments of incandescent light bulbs. The actual

r

march of this newly revived technology began inbetween the two

IIT py

world wars and after that it has never looked back. Today powder

metallurgy parts range from light metals to heavy metals, low

o

melting materials to refractory materials and from highly porous

parts to dense structural parts.

C

An unique attribute of powder metallurgy is the a b ilit y

to produce porous products with interconnected porosity. This

attribute, according to the Sourse Book on Copper and Copper

Alloys1 , made possible the development of the self-lubricating

bronze bearings, an early P/M product, the f ir s t having been

used in a Buik automobile in the 1 9 2 0 ’ s. Depending on the

sintered density these bearings can absorb from 10 to 30$ by

volume of oil and can supply a continous lu b r ica tin g film even
 2
at low speeds. Porous bronze bearings also have the advantage

that they are sufficiently ductile to permit assembly by ring

staking.

Self-lubricating bronze bearings depend on conduction

and convection for heat dissipation during service. According

to Morgan the idea of self-lubricating bearings probably

originated from the attempts to overcome the heat conductivity

limitation of oil soaked wooden bearings. The frictional heat

r
developed is proportional to PVp. where P is the pressure on

pu
the bearing, Y is the surface velocity and p, is the
ar t
coefficient of friction.
h
ag Although the automobile industry, which was responsible
Kh ig
for the development of these bearings, still uses them extensively

their applications have now spread over to many other industries.

r

Development of these bearings revolutionized the home appliance

IIT py

industry. By eliminating the requirement of periodic lubrication,

the self-lubricating bearing assured many years of trouble-free

o

operation of home appliances and led to a great expansion of

the industry. New applications continue to be found and the

C

self-lubricating bronze bearing industry consumes a major portion

of the copper powder produced each year. According to a review

3
by Taubenblat the manufacture of porous self-lubricating

bronze bearings accounts for about 75 % of the granular copper

powder consumed by the industries.

Self-lubricating porous bronze bearings are used in

large numbers in fractional horsepower electric motors.

 3
Applications in which such small motors are used has been
4
given by Victor . Some of them are as follows t electric fans

and blowers, sewing machines, food mixers, refrigerators, air-

conditioning units, textile equipments, automobile heaters and

defrosters, windshield wipers, e .t .c . The list is unending and

is well summarized by Tait

5 in a broad review of non-ferrous

powder metallurgy applications in automobile industries.

Although bronze is found to be an excellent bearing

r
material and is extensively used in industries, iron bearings

pu
are finding wide applications recently. To a large extent, the
ar t
world wide copper shortage has contributed to this development.
h
Iron bearings have other advantages in lower material cost and
ag
Kh ig
higher strength but rust very quickly. So nearly in all cases

where long, trouble-free bearing life is required, bronze

r

bearings are the best.

IIT py

A small group of special types of bearings, suited for

very special applications have been outlined by various

o

authors • Some of them are high temperature bearings,

gas-bearings, aluminium bearings, steel backed bearings,

C

copper-lead bearings, lead-infiltrated bearings, dry bearings

e .t .c . It must be remembered that these bearings are in no way

an alternative to the normal bronze bearings. These special

bearings are used only in places where the normal self-

lubricating bronze bearings cannot be used owing to certain

service conditions. As a matter of fact, with each passing year

the demand for self lubricating bronze bearings is continually

 4
increasing, the annual world production being estimated to be

about eight million per day in the year 1964, and is still

expanding.

Though the self-lubricating bronze bearings are being

manufactured from a very long time back, many of its manu­

facturing parameters are very vaguely given in the literature.

The correlation of manufacturing parameters with the bearing

properties are not very clear from the available literature, and

as for the physics of sintering is concerned there are two

r
pu
distinctly contradicting theories.
ar t
The general outline of the manufacturing process of
h
self-lubricating bronze bearings has been described by a number
ag
Kh ig
1 2 -2 0
of authors . A good comparison between an old method of

bearing production with the powder metallurgy route is given by

r

20
Borden . According to his article * ‘Previously bushings had
IIT py

to be turned and bored from brass bar stocks on automatic screw

machines. Production averaged about 230 pieces per hour by this

method with rejects often as high as 50$ due to dimensional

o

inaccuracies. The same bushing, of sintered bronze, is now being

C

turned out at a rate of 4000 per hour by the briquetting press

and 12,000 per hour by the furnace under full production, with

practically no rejects* .

The whole process can be broken up into a number of

steps like mixing, compacting, pre-sintering and sintering,

repressing, oil impregnation and packing. These steps has been

2
outlined in short by Morgan '
 5
"M etal powders are mixed to given the desired alloy

composition, for example, 89$ of copper powder, 10$ tin powder

and 1% of powdered graphite are mixed to produce a graphited

bronze.

The mixed powders are compacted to the desired shape in

precision tools on automatic presses at pressures between 10

2
and 35 tons/inch , . The compacted briquettes are sufficiently

strong to permit handling.

r
The green compacts are sintered in a reducing atmosphere.

pu
Before sintering it is generally pre-sintered at lower
ar t o
h
temperatures of about 400 C in order to drive out the lubricants

and oils. For 90/10 bronze a sintering temperature of about

ag
Kh ig
800°C is used.
r

The sintered compacts are repressed to remove distor­

IIT py

tion and to impart dimensional accuracy and a smooth surface

fin ish .
o

The pores of the coined components are filled with

lubricating oil by vacuum impregnation.

C

The finished bearings are packed in a non-absorbent

wrapping, to protect them from dust and avoid loss of o il1

,’

The steps briefly elucidated above may be considered

to be the process variables in the manufacture of self-

lubricating bronze bearings. Another important variable which

influences the properties of self-lubricating bronze bearings

 6
is the material variable. Amongst the material variables one

can consider the characteristics of copper, tin and graphite

powders, their amounts, and also the amount and type of

lubricants used. The type of material and process variable used

w ill determine the final properties of the self-lubricating

bronze bearings. So it would be best to discuss the properties

of self-lubricating bronze bearings under the two broad heads

of material and process variables.

r
1 .1 MATERIAL VARIABLES

pu
ar th The main material variables which affects the final

properties of self-lubricating bronze bearings has been

ag
Kh ig
outlined below.

1 .1 .1 Copper
r
IIT py

Copper, which happens to be the major elemental

constituent of the sintered self-lubricating bronze bearings

can be manufactured by a number of ways. The production and

o

properties of copper powders have been discussed by a number

11 21-27
of authors * . In general the routes of manufacturing
C

which are followed includes electrolysis, hydrometallurgy,

gaseous reduction and atomization.

The properties of the electrolytic copper powder

deposited from sulphate bath depends on the variables that

effects the electrolytic process. The oxygen content of

electrolytic copper powder is higher when compared with other

 methods. This oxygen content is generally seen as a thin

surface film of cupric oxide giving a dark brown to black

colour to the powders even when the oxygen content is less than

0 ,5 % . Occationally, the oxide film is either cuprous oxide

or a mixture of cuprous and cupric oxides. Rarely particles

28
having a surface of CU4 O have also been found

This type of powder is s t ill used to some extent in the

manufacture of porous bronze bearings. Its high cost of

production, however, makes it a less attractive proposal for

r
manufacture of porous bearings and is presently being replaced

pu
ar t
by reduced powders and when possible by atomized powders.
h
The method of gaseous reduction generally relies on
ag
Kh ig
copper mill scales as the starting material. Also copper oxides

obtained by the oxidation of chemical precipitates, or finely

r

divided electrolytic copper powders may sometimes serve as

IIT py

the feed material. The oxides are comminuted and reduced to

the metal by using carbon monoxide or hydrogen.

o

Mahindra Sintered Products, one of the largest manu­

facturers of self-lubricating bronze bearings in this country,

C

uses reduced copper powder. The powders produced by the

reduction of mill scales, generally exhibits good green

strengths and high apparent densities. It is generally spongy

in nature. Also its manufacturing cost is not as high as that

of electrolytic copper powder.

As a method of comminution, probably the cheapest

possible process is that of atomization. The fin a l properties

 8

of the powders is greatly influenced by the raw material used,

the atomizing medium and the final processing of the powders.

The over-all oxygen content is not excessive, since the

particles are substantially solid and spherical or tear drop

shaped, and the films are generally so thin that they are

literally transparent. Though this powder has excellent flow-

ability and cheap manufacturing cost, owing to its extemely

bad pressing properties, it was for long not used by the

bearing industries.

r
pu
In recent years, considerable flexibility and control
ar t
has been achieved in metal atomization precesses and it is now
h
possible to make copper powders having highly irregular
ag
Kh ig
particle shapes with apparent densities as low as 2 .1 g /c .c .

Though these powders are in the simple atomised condition,

r

their physical appearence is remarkably similar to that of

IIT py

reduced copper powders.

1 .1 .2 Tin
o

Tin is the next important metal required in the

C

manufacture of self-lubricating bronze bearings. A great deal

o f literature and patent survey of this metal was done by

29
Watkins . Tin is generally produced by a number of methods

like shotting, graining, atomization, electrolysis and

chemical precipitation . Generally for the powder metallurgy

purpose the last three is of use as it yields tin fine enough

to be used for mixing and molding work.

 9
The most commonly used powder for the manufacture of

self-lubricating bronze bearings is atomised tin. According

30
to Goetzel atomization results in a fine product whose

particle size distribution can be regulated over a wide range.

The oxygen content of the atomized tin is normally below 0 .2 $ .

By varying the atomizing parameters the partiele sizes can

also be varied.

31
According to Elliot previously self-lubricating

bronze bearings were made with a wide range of compositions,

r
the tin content varying between 5 to 15$ by weight. Recently

pu
ar t
however, there is a tendency to standardise around
h 10$

compositions. Standards like ASTM, SAE, MPIF also stipulate

tin

ag
a composition of around 10$ tin , Tin generally strengthens
Kh ig
1
the copper by solid solution hardening. So with increasing
r

tin percentages the strength of the bearings w ill increase.

IIT py

1 .1 .3 Lubricants

Generally non-metallic additions are done for many

o

reasons. They may be classified as follows Transitory

additions which disappear during the sintering cycle or

C

permanent additions which are incorporated in the final

product. This latter mentioned eatagory may f u l f i l l any of

the following purposes a) form an alloy composition

b ) act as free ingrediants c) promote controlled porosity

d) act as binders e) act as lubricant.

In self-lubricating bronze bearing production,

generally both types of additions are made. One of the

 10
lubricant which is transitory in nature is added to reduce

the friction between powder and die wall and the other is a

permanent addition which is to provide lubrication during

operation of the bearing.

In order to decrease the frictional forces between the

powder and die wall, lubricants are either applied to the die

wall or admixed with the metal powder. Though lubrication plays

an important role in the production of parts by powder

metallurgy, very little is known about the best way to apply

r
32
lubrication . A number of authors have expressed various

pu
ar t 33"“39
views as to the type and amount of lubrication to be used
h
Generally in industries only upto 1% by weight of these

ag
transitory lubricants are added.
Kh ig
Admixed lubrication is suitable for automatic presses.
r

Since there is high speed movement the compound preferred

IIT py

should have high melting point and a long chain. These

lubricants are driven out during the sintering cycle.

o

The other type of non-metallic addition that is used

in the manufacturing process of self-lubricating bronze bear­

C

ings is a permanent addition. This is added to help in the

lubrication of the bearing during its operation. Generally

graphite is added for this purpose, though presently some

special bearings for use in space vehicles has added to it,

Molybdenum disulphide as the solid lubricant. These graphite

or molybdenum disulphide which are added do not basically

affect relations between the metallic components, acting as an

 11
interfering phase in the boundary region, similar to a pore.

31
According to Elliot previously self-lubricating

bronze bearings contained graphite of upto even 6% by weight,

but recently graphite content is being limited to a maximum of

38
2% . More recently Berry has stipulated that specifications

for bronze bearings permits a graphite addition upto a maximum

of 1,5% . Low graphite contents are naturally preferred as

graphite generally reduces the strength of the bearings. Bronze

r
bearings are also made without the additions of any graphite.

pu
ar thAnother addition that is generally made before the

mixing stage is the light o il. This oil is generally used to

help and prevent the finer tin powders from segregating.

ag
Kh ig
However, addition of excess light o il to the powder decreases
40
the flowability. Kosho in his work on bronze bearings used
r

light oil upto 0 ,2 % (weight ra tio ). This is removed by heating

IIT py

o
before the sintering cycle for 60 minutes at 150 C in hydrogen.

1 .1 .4 Particle sizes
o

The powder particle size distribution plays an

C

important role in influencing the final properties of self-

lubricating bronze bearings. In reality there is no fixed size

distribution specified for manufacture. It varies from user to

user and is determined by considering the buyers demand and

compromising between conflicting requirements. According to

41
Kieffer and Hotop the powder grades may vary from minus 50
■/
mesh to minus 325 mesh. According to Jones the tin powder size
 12
is of less importance. Hall indicates that tin powders of

considerable fineness are extensively used by bearing

manufacturers. Some size distribution of commonly used copper,

tin and graphite powders used for the manufacture of self-

13 12
lubricating bronze bearings are given by Goetzel .J o n e s ,

in the manufacturing process of porous bronze bearings, said

that copper powder used is finer than 100 mesh out of which

around 50-60$ passes through 325 mesh and the tin powder used

is generally finer than the copper powder.

r
43

pu
Hayashi and Watanabe in their works, used two types
ar t
of tin powders. One contains 60$ by weight of 100-150 mesh
h
powder and 40$ by weight of minus 270 mesh powder, and the
ag
Kh ig
other is totally minus 270 mesh. Their work shows that as the

temperature increases beyond the melting point of tin, the tin

r

melts and gets absorbed between the copper particles. The

IIT py

voids left by the tin particles become pores. The tin gradually
o
diffused into copper. After pre-sintering at 400 C, when the
o
temperature is being raised to 800 C, the diffusion becomes
o

o
active. When the temperature reaches 750 C alpha phase begins
o
C

to appear. When samples are held for 20 minutes at 750 C new

crack like pores appear in addition to the earlier ones.

These capillary type of pores are formed due to the sticking.

together of the particles and their resulting movement.

However, when the samples are maintained for 60 to 180 minutes

o
at 750 C the pores become spherical and with longer holding

times the samples shrink and pores become smaller. However'

 when sintering fine tin powder, above the melting point of

tin, tin undergoes only diffusion and does not ooze out into

the space between the copper particles and hence new pores
o
are not formed. However, as temperature reaches 600 C and

alpha phase begins to form, pores are also formed. In samples

o
maintained at 750 C for 60 minutes pores have also increased

and they are crack like ones. The authors belive that such

pores are formed as follows as sintering progresses the alloy

phase tends towards the copper side of the equilibrium diagram.

r
The copper-tin alloys other than the alpha phase are fragile

pu
ar t
and have lower melting points. So a part of the sintered body
h
o
is in the molten state at 750 C . Now the sample is undergoing

shrinkage due to the sticking together of the particles. The

ag
Kh ig
shrunken regions, as a result of formation of alpha phase,

becomes seperated from the non-alpha region and the space

r
IIT py

between the two regions becomes crack like pores. Usually when
44
a porous substance shrinks the pores are decreased but here

the pores increase due to the above mentioned reason. Even­

o

tually, however, overall shrinkage occurs reducing the porosity,

40
C

Kosho studied the importance of tin particle size

in controlling the final properties of self-lubricating bronze

bearings. According to him the pores develop due to a 11 skin

shedding11 phenomenon of tin powder, and concludes that pores

can be controlled with the help of tin powders. He studied the

effect of the difference in granularity of tin powder on size

and number of oil holes. In case of a sintered body, contract­

ion becomes more intense as the tin powder becomes finer. The
 tin size becoming finer means the surface area of the powder

increases and its distribution in copper powder becomes

finer. So the alloying of copper with tin becomes faster and

a strong tendency of densification is observed. Also the pores

become finer during the sintering process and they contract

and become isolated. With increasing fineness of tin the

hardness and density of the compacts also increase. As far as

the grain size of the sintered specimens are concerned, they

become finer with increasing fineness of the tin powders.

r
Another observation is that keeping the amount of tin same

pu
( 10$ by weight) and varying the granularity of the tin powder,
ar th
the number of pores in a particular place increases geometri­

cally. Hence the conclusion is drawn that by varying the

ag
Kh ig
granularity of the tin powder, control can be exercised over

the size and number of pores.

r
IIT py

The reason for the grain size becoming finer with

45
finer tin powders has been explained by Kosho in a later

article. According to him if fine grained tin powder is used,

o

hole formed by the ’* skin-shedding” of the tin powders are

C

finely distributed and as a result growth of crystal grains

are hampered.

46
Kosho in another work studied the effect of varying

tin size on the oil permeability and oil retention capacity.

In samples with coarse tin powders which gives coarse oil

holes the oil permeability is very good but in contrast to

this, since the capillary force is weak, the oil retention

 15
capacity is very weak. The opposite effect is seen in case

of samples with fine oil holes.

The effect of close porosity control and proper

selection of the lubricating oil to be used is stressed by

47
Geotzel . The pore size can be regulated readily by the

grain size of powders used, which can range between 0.0015 inch

(0 .0 0 3 8 cm) to 0.006 inch (0.015 cm). Bearings to be subjected

to heavy loads at high temperatures require a heavy oil

r
(SAE 60) . In its manufacture a comparitively coarse powder

pu
must be used with smallest particle size not below 0.004 inch
ar t
(0 .0 1 cm) in diameter. For bearing to be used for high speeds,
h
low loads and low temperatures, a light oil with low viscosity
ag
Kh ig
(SAE 20) is required. Its manufacture should be from fine

powders, 0.002 inch (0.005 cm) being the smallest particle

r

48
diameter. Also according to Pratt close control of the
IIT py

powder size fraction and degree of compaction gives the

49
required level of porosity. Jones has reported a Graphite

free porous bronze material containing 9 0 .5 $ copper and 9 ,5 %

o

tin as having very large pores (0 .7 5 mm) which is claimed to

C

be suitable for high load applications requiring external oil

supply, without reducing the capillarity of the material

required for the reabsorption of the o il. The large pore size

acts as a factor which aids in the distribution of o il, which

would have been normally impeded by opposing suction forces.

50
Mazzarella in his work used nine types of copper

sizes and he observed that the porosity decreased with the

 16
powder fineness, which is to be expected.

51
Lennox and Brewer in their article on porous

bronze bearings have said that powder used must be rigidly

held to the specification agreed upon by powder manufacturer

and user. In the production of porous bronze bearings it is

usual to have an electrolytic copper powder and atomized tin

powder. Considerable care must be taken to prevent oxidation

before use and this necessitates transport in hermetically

sealed containers.

r
pu
ar t In general it can be concluded that increasing the
h
fineness of both copper and tin powders, the rate of

homogenization becomes faster. So dens i f i cation, hardness and

ag
Kh ig
radial crushing strength increases with the increase in fineness

of both the powders. Also with increasing fineness of the

r

powders the total porosity and also the interconnected porosity

IIT py

drops. The lubricants that are added to the powders generally

decreases the strength of the bearings. The graphite amongst

o

the lubricants is known to cause growth of the compacts during

sintering.
C

1 .2 PROCESS VARIABLES

1 .2 .1 Mixing

The elemental powders are blended in ball mills or

tumbler m ills. The mixing is to be regulated carefully as

excess time of mixing w ill result in the breaking of the

 17
larger powder particles thus giving a overall change in the

particle size distribution. On the other hand i f the mixing

is carried out for too short a time than the tin, not being

uniformly distributed in the copper will give irregular

40
properties. It was observed by Kosho that the addition of

light oil . is a must in order to maintain a homogeneous

distribution of tin in copper.

1 .2 .2 Compacting

r
13

pu
ar t According to Goetzel the compacting is generally

carried out in quick acting mechanical presses. Hydraulic

h
presses may also be used for compaction. The pressures of
ag
Kh ig
briquetting varies considerably with the individual manu­

facturer and also the type of bearing that is to be made.

r

Generally the pressures used are within the range of 15 to

IIT py

35 tsi (235 to 550 MPa), but it may also be as low as 5 tsi

(7 8 .5 MPa) or as high as 50 tsi (785 MPa) when an ultimate

sizing operation is to be eliminated. Depending on the size

o

and shape of the bearings the compaction rate may vary from
C

six to several hundred compacts per minute.

40
Kosho used moulding pressures from 0 .6 to
2
1 .6 T/cm (5 9 .7 to 1 5 9 ,3 MPa) and also varied the particle

size of tin in his mixes. He found that the green density

increases almost linearly with the increase in pressure. The

sintered density of the compacts increased with the green

 density. High sintered density compacts exibit more closed

porosity. An increase in the compacting pressures also

results in the increase of the mechanical properties like

the hardness and the radial crushing strength.

38
Berry m his works on factors affecting the growth

of 90$ copper and 10$ tin mixes pointed out that the growth

increases as green density increases ( i .e . compacting

pressure increases). His reasons are that with increasing

r
green density the available surface are of copper for alloying

pu
with tin w ill decrease and so an increase in growth can be
ar t
expected. Also at high green densities the swelling can be
h
more owing to pressure build up during delubrication or the
ag
Kh ig
escape of other gases from the compact.
r

1 .2 .3 Sintering
IIT py

The whole process of sintering can be broken up into

a number of seperate sub-heads and then treated. The main

o

sintering variables are the sintering temperature, sintering

time, the rate of heating up to the sintering temperature and

C

the sintering atmosphere.

1 .2 .3 .a Sintering temperature! In this step the pre­

sintering may also be considered. The general pre-sintering

treatment is carried out in the temperature range of 400 to

o 13
500 C. According to Goetzel this pre-sintering serves the

twin purpose of expelling the volatile lubricants at a low

 rate, and also permits sufficient time for the liquification

o f the tin and its diffusion into copper.

Most of the authors refer to the sintering tem-

o • 12
perature, loosely, as around 800 C • According to Jones
o
the sintering temperature is between 800 to 850 C. Amongst
52
the early workers, Koehring observed that a sudden swelling
o
occured on sintering at 800 C . Similar increase in growth
53 o
is also observed by Mitani at 798 C in his dialetometric

r
work, and the temperature coincided with the first peritectic
54

pu
reaction temperature. Esper and Zeller in their dialeto­
ar t
metric works found breaks in the expansion curves at 230, 450,
o
h
640, 750 C which is said to correspond to new liquid phases
ag 31
Kh ig
formed. Elliot in studying the influence of sintering

temperature on growth in the radial direction observed that

o
r

from 700 G the radial growth continually increases t ill a

o
IIT py

temperature of about 825 C is reached and then it drops. He

however did not find any significant growth increase as soon

o
as the first peritectic reaction temperature at 799 C is
o

crossed. He attributed the growth factor with temperature to

C

be associated with the progressive fusion of the various

intermediate solid solutions,

45
However Kosho in his works on influence of sinter­

ing temperature on some properties of sintered bronze bearings

found that there is not much change (even when the sintering
o
temperature crosses 798 C) observed in the linear contraction

and sintered density for mixes having coarser tin powder*

 20
With the increase in the fineness of the tin powder the

linear contraction and the sintered density increase with

increasing sintering temperature. This is however contra­

dictory to the findings of most of the other workers. Also

according to his work the interconnected porosity when

expressed as a percent of total porosity decreases when the

o
sintering temperature is taken from 790 to 810 G.

To explain the changes in properties of self-lubricating

r
bronze bearings with the rise in sintering temperature, it is

pu
important to look into the structure formation mechanism in
ar t
the sintering of such bearings. According to Geotzel
55 , after
h
the melting point of tin at 232°C is crossed the tin melts,
ag
Kh ig
and with increasing temperature and time the molten tin

diffuses rapidly into the surrounding copper particles so that

r

after the sintering temperature is reached and the compact

IIT py

remains there for a length of time, a homogeneous solid

solution of alpha bronze is formed. According to colour photo-

4 2 i— 55
micrographs taken by Hall |[_as reported by Goetzel J
o

actual phase changes that occur during sintering are shown.

C

Diffusion between tin and copper is said to commence after one

minute at the temperature where tin-rich transitory phases

are formed. After around 25 minutes, a homogeneous alloy of

alpha-bronze is obtained. The X-ray corroboration with the

56
microstructural properties are given by Carter and Metelfe .

According to them the porosity and the phase distribution of

the material is sharply influenced by the peritectic reaction

 21

o 55
temperature at 798 C. According to Goetzel the schematic

representation of the sintering mechanism has been given by

57
Saurwald .

12
According to Jones , the mechanism of structure

formation of bronze bearings during sintering of elemental

powder mixes is quite complex. According to the copper-tin

phase diagram a number of intermetallic phases may make their

appearence with increasing temperature, either at the interface

r
or in the tin melt. According to him the more important of

pu
these are first
ar t V (approximately 60 % tin) at low temperatures,
o
h
followed by c (approximately 38$ tin) stable upto 670 C, and

then finally at higher temperatures Y and £ and then the

ag
Kh ig
copper-tin solid solution °( . Finished sintered bronzes

generally contain either only «< (which may or may not be

r

homogeneously diffused) or sometimes a mixture of °C and 5

IIT py

produced by the decomposition of the £ and Y phases. In

this case it can be assumed that the diffusion process is not

complete and the sintering does not proceed to the complete

o

formation of <*, but leaves behind some £ phase. According

C

to him there is some divergence of opinion as to whether the

residual 8 phase is beneficial or not. The rate of solution-

ing of solid copper in liquid tin is very high, and it is

probable that, within seconds of the melting of tin the first

copper particles (less than 15 microns) are completely

dissolved. The compact has excess of copper so the rate of

solutioning quickly decreases as copper content of molten tin

 rises. Solution of copper in the tin with the formation of

the V and then the c phases proceeds rapidly and unless

the heating up is very rapid the compact is wholly solid by

o
the time 500 C is attained. Then the microstructure should

have copper particles, with some enveloping c< phase formed

by diffusion of tin into the copper lattice, surrounded by

zones of 5 and e phases. He agrees that the course of the

subsequent diffusion is not known with certainty, but he says

that probably tin diffuses from these phases into the copper

r
lattice, though the reverse process could also be possible.

pu
ar t
When sintering temperature is taken above the first pertectic
h
o
reaction temperature at 798 C, any residual £ phase liquifi

and the homogenization process is speeded up.

ag
Kh ig
The pores consists partly of the previously existing
r

pores in the compact along with some of the volume that was
IIT py

previously occupied by the tin particles. The copper particle

however expands because of diffusion of tin into them. Also

the melting ‘ of tin allows stress releiving and repacking of

o

the particles, and probably the tin moves to zones of least

C

pressure. The graphite only acts as a barrier to the fast

diffusion process.

It is generally agreed that the growth observed in

the bronze compacts depands on the alloying of tin into

38
copper. According to Berry , on melting the tin tends to

alloy in the regions of copper particle junction, the

increase in volume produced by diffusion causing the copper

 23
particles to move apart and growth to occur. While alloying

is producing an increase in volume, there is an opposing

sintering process in action reducing the surface area of the

pores. Diffusion must he considered a two way process, and in

the case of the copper tin system the diffusion of tin in

50
copper is greater than that of copper in tin , Asaeda

maintains that growth occurs in a copper based system where

diffusion in copper is greater than in additives, while

shrinkage is favoured if the diffusion in the additives is

r
greater than that in copper,

pu
ar t 59
Daver and Ullrich carried out some work on the
h
structure formation mechanism in copper tin systems (90 copper
ag
Kh ig
10 tin) commonly used for sintered bronze bearings. The work

was undertaken to study the liquid phase sintering with the

r

help of hot-stage microscope. Their findings in a nutshell

IIT py

describes the theory of structure formation mechanism of

90 copper - 10 tin compacts. According to them, as soon as the

tin melts, it reacts with the copper particles forming

o

intermediate phases - first V (60 % tin) followed by e (38$

C

t i n ) . The phases are formed at the interface and not in the

tin melt. The areas originally occupied by the tin phase

leaves behind pores of approximately the same size and shape.

At around 700°C crystal formation and metallurgical bonds

develop rapidly. The phases that now exist are and 5 . The

5 phase is retained at the room temperature (when cooled) as

it decomposes to <=( plus e. only under complete equilibrium

 conditions. When 800 C .is crossed diffusion increases

tremendously accompanied by concentration leveling, homo-

genization and solidification. The structure is °( bronze with

evidence of twinning, pore migration and spherodization.

60
Znatokova and Likhtman agree that from the experi­

mental data obtained by them, they have not catagorically

proved the structure formation mechanism. They have pointed

out that existing theories relating to the mechanism of

r
structural transformations in the copper tin systems are

pu
contradictory.
ar th
According to them, during the sintering of a copper

tin mixture the constituents diffuse into each other leading

ag
Kh ig
to the formation of intermetallic compounds, and in the long

run, to the formation of a homogeneous solid solution of tin

r

and copper. In their model experiments carried out at 200 -

IIT py

o
250 C they found that it is copper which diffused into tin

forming new phases on the copper tin boundary and within the
o

volume of tin. When sinterings are carried out at temperatures

below the melting point of tin a transitional diffusion layer

C

is found, consisting of two phases, e phase bordering upon

the copper and V phase bordering upon the tin . When the
o
temperature is taken above the melting point of tin (232 C ),

the transitional layer on the tin side has on uneven edge

with stretched and often torn crystals of V phase. The V

phase is brilliant white and the c phase is gray in colour.

These new phases have different microbardnesses from that of

 t in . Microhardnesses of tin, V (white phase), and c

(gray phase) are 16-25, 460-610, 179-285 Kg/mm2 respectively.

On a further rise in temperature the process develops

in the direction of an increase in the volume of c phase,

both through the decrease in the volume of phase and

V
o
through the surrounding copper. On heating upto 400 C, all

traces of 1? phase disappears from the areas rich in tin

which then consists of entirely c phase. Then on heating

o

r
to 420 C, on the periphery of these areas there appears a

pu
thin layer of a new phase called
ar t 5 phase, which is still

richer in copper and also white in colour like the V phase.

h
With increasing temperature this layer extends deeper absorb-
ag
Kh ig
o
ing the c phase and at 500 C this process is nearly

complete. In addition to this, in the temperature region of

r

o
400 - 500 C there is extensive «=c solid solution formation
IIT py

through diffusion of tin and copper, accompanied by breakdown

of the intermetallic compounds that have formed, and this

o
process is nearly complete at 600 C, At this temperature solid
o

solution forms every where with traces of crushed <=(-5

C

o
eutectoid. At 600 C the solid solution which was hetero-
2
geneous (microhardness 165-233 kg/mm ) becomes nearly homo-
o
geneous at 700 C with the 8 phase being restricted to very
o
small areas. At 800 C a homogeneous solid solution is
2
obtained (hardness 213-233 kg/mm ) •

Znatokova and Likhtman have compared their results

with the other published ones12* ^ According to

 them,many authors are of the openion that copper tin system

should be regarded as a system containing a liquid phase

within a wide temperature interval. After melting tin is

drawn by capillary forces into micropores between the copper

particles and it spreads over the surface of these particles

and envelops them in a thick film. Furthermore, with increas­

ing sintering temperature the process of diffusion penetration

of tin into copper is intensified and leads to the formation

of new phases and finally to the formation of homogeneous °(

so lid solution. However, from the work of Znatokova and

r
SO

pu
Likhtman the process of sintering of the copper tin system
ar t
cannot be considered as taking place in the presence of a
h
liq uid phase. Even if heating is sufficiently rapid for the
ag
Kh ig
tin to remain unaffected upto the melting point, then the

liquid phase forming is immediately enveloped in a thin but

r

rapidly thickening layer of solid phase, arising as a

IIT py

result of diffusion of copper in tin, and preventing the tin

from spreading. Thus it is not possible for liquid phase to

remain for an extensive length of time at temperatures above

o

232°C, since V phase disappears at 400°C, i . e . below its

C

melting point, and is replaced by the more refractory c

and then by 6 phase . The latter also breaks down, forming

o< solid solution at 600°C, i . e . again below its melting point.

12
1 . 2 .3 . b Sintering time : According to Jones an important

point that is to be noticed, but is not realized, is that the

whole thermal cycle is usually very short. Time varies with

 different manufacturers, but it is quite eiistomary for the

temperature of the compacts to be raised to the sintering

temperature in 2 to 5 minutes, and for the final sintering

treatment to last no more than 3 to 5 minutes. Cooling

generally arranged by water jackets around the furnace muffle,

o
takes longer, but the compacts reach 100 C in less than 10
42
minutes. In his work, Hall found that with 25 to 27 minutes

at the sintering temperature the alloying is substantially

completed, as evident from the appearance of <=c bronze.

r
pu
64
Kosho
ar t , in his detailed work, studied the final

properties as a function of sintering times (0 to 240 minutes)

h
and of the tin particle size and of sintering temperatures
ag
Kh ig
o o o
of 790 G and 810 C . At 790 C, porosity gradually decreases
o
with sintering time. When sintering at 810 C, the porosities
r

o
are lower when compared to the ones sintered at 790 G. At
IIT py

810°C also, with sintering time the porosity decreases.

Finally, he also observed that the tendency of contraction

of oil holes is there with the contraction of sintered body

o

o
when held at 790 C - but closed pores are not formed. However,
C

o
when the temperature is held at 810 C, more and more closed

pores forms with sintering times.

65
In the article by Taubenblat and Smith , 90/10
o
bronze bearings are sintered at 843 C for 3 ,5 and 13 minutes.

According to the microstructures taken at the three times it

is seen that the shortest time of sintering produced an

 28

undersintered structure consisting of uneven grain sizes

with particles of unalloyed copper remaining. The K-factor

of this is found to be around 18,000 psi. Sintering for 5

minutes gives good optimum structure consisting of a single

phase matrix, with small and uniform grain size, and a K -

factor of 31,000 psi. However when sintering for 13 minutes

the tin is completely diffused into copper which results in

a single phase structure, but with excessive grain growth.

The grain sizes are bigger when compared to the samples

r
sintered for 5 minutes and it gives a K-factor of 29,000 psi

pu
ar t
which is lower than the optimum sintered one.
h
66
May , studied the electrical resistance during
ag
Kh ig
inductive liquid phase sintering of copper-tin mixes. He

observed that the resistance drops as the sintering tempera­

r

ture increases from room temperature to the melting point of

IIT py

t in . As soon as the melting point of tin is crossed there

is a sudden rise in the resistance value - followed by a

o
continous decrease t i l l about 800 C is reached. As soon as
o

a temperature of 800°C (which corresponds to the first

C

peritectic reaction temperature at 799°C) is crossed there is

again a sharp increase in the resistance value t i l l the point

at which the maximum sintering temperature is reached. It is

seen that on holding at the maximum sintering temperature,

the resistance value gradujally decreases with time, and

then with s t i l l further holding passes through a hump. The

author has. however failed to give much explanation for the

nature of the curves.

f
 Elliot
31 working on some factors which affects the

growth of copper-tin compacts, showed that the samples

o
reached 800 C in about three minutes. On reaching the

sintering temperature, the growth remained fairly constant

for several minutes, and after ten minutes a slight reduction

in overall growth occured. Following the grow/th which arises

during the in itia l period required to reach the sintering

temperature, it seems that the porous structure which is

established is dimensionally stable. The increase in the

r
shrinkage factor, which could be expected to develop with time

pu
did not manifest itself until the end of the sintering cycle.
ar th
38
Berry*3 also working on effects of sintering time

on the growth of 90 copper / 10 tin compacts concludes, that

ag
Kh ig
holding at temperatures for times much longer than 5 minutes

does not appreciably reduce growth under the sintering

r

conditions of the test.

IIT py

1 .2 .S .C Heating rate i That heating rate w ill affect the

o

fin a l sintered properties of bronze bearings is one of the

12
conclusions drawn up by Jones . According to him it w ill
C

affect the time upto which a part of the compact remains in

the liquid state. " Solution of copper in tin with the

formation of the V and then the c phases w ill rapidly

proceed, and very probably, unless the rate of heating up is

very rapid, the compact is wholly solid again by the time

o
500 C has been attained" • However, according to Znatokova
60
and Likhtman the process of sintering of the copper-tin
 system cannot be considered as taking place in the presence;

of a liquid phase, even if heating rate is sufficiently rapid

for the tin to remain unaffected upto its melting point. The

liquid phase forming is immediately enveloped in a thin but

rapidly thickening layer of <=( phase, thus preventing the

liq uid phase from speading.

31
Work done by Elliot shows that in conveyor furnace

the growth is 2 .2 5 % whereas in batch furnace it is observed

r
to be 3% . This increase can be attributed to the higher rate

pu
of heating in batch furnace, which has a substantial influence
ar t 38
upon growth. Berry also working on growth of copper-tin
h
compacts agrees that under rapid heating rate that prevails in
ag
Kh ig
practice, it is likely that some liquid phase is present

throughout sintering until the final solification above the

r

peritectic reaction temperature occurs. According to him the

IIT py

growth increases as the time to reach the maximum temperature

is decreased. Increase in heating rate effectively reduces the

extent of tin diffusion before the peritectic temperature is

o

reached. More liquid phase is thus produced and causes a

C

greater growth of the compacts.

Detailed work on effect of heating rate on the

properties of 90 copper / 10 tin bronze bearings is reported

64
by Kosho . In his work, the presintering temperature used

is 400°C, heated upto it at the rate of 15°C / minute, and

held there for 60 minutes. Then from 400°C to 790 C the heat­
ing rates are varied from V 3 to 80°C / minute. He observed
 31
that porosity increases with an increase in the heating rate,

but the interconnected porosity is not much affected by the

increase in porosity. Very poor results are obtained by

o
heating at the rate of 3 C / minute but a rapid improvement

in the interconnected porosity is noticed with an increase in

the heating rate. Almost a normal state of 85$. of total

porosity is reached when heating rate is 15°C / minute, and it

remains constant above this heating rate level. Samples heated

o
at the rate of 80 C / minute shows unusual distortion. From

r
microstructural observations he noted that with the increase

pu
ar t
in the heating rates the globular pores grows larger.
h
13
1 .2 .3 . d
ag Sintering atmosphere *. According to Goetzel ,
Kh ig
11 Sintering is also exclusively conducted in continous furnace

in a protective atmosphere, usually consisting of partially

r

combusted hydrocarbons or natural gas, but in certain instances

IIT py

protection against oxidation during sintering is achieved by

simply packing briquettes together with coke or charcoal

o

inside closed metal boxes. Hydrogen and Dissociated Ammonia

are not generally used because of the higher costs involved" •

C

Studies on effect of atmosphere on properties of

53
bronze bearings are few . However Mitani working on

dialatometry, obtained an abnormal expansion when the

temperature crossed the first peritectic reaction. He con­

sidered that the abnormal growth can be reduced by using an

atmosphere containing a minimal hydrogen content, since his

findings indicate that a significant proportion of the

 expansion is caused by evolution of dissolved hydrogen on

solidification.

38
Berry and Meyer et al has shown that decomposi­

tion of lubricants are influenced by the atmosphere. Berry28

showed that a negative growth can be produced i f the volatiles

on delubrication are of a reducing nature, particularly where

lean atmospheres are used.

The influence of sintering atmosphere on the pores are

68

r
briefly described by Fleming . It is known that sintering

pu
affects the porosity in the following manner a) the total
ar t
number of pores decreases progressively during sintering at
h
a ll temperatures (b) the average pore size existing in the
ag
Kh ig
compact increases with the sintering time, (c) there is, for

each conditions of sintering, a pore size which occurs in

r

maximum number in the size scale, and this maximum shifts

IIT py

towards larger pore size as sintering proceeds, (d) late in

the sintering process there exist pores which are larger than

any present at the beginning of sintering.

o

These factors are independent of whether sintering

C

occurs in a reducing or neutral atmosphere, or in vacuum. The

affects of vacuum sintering or sintering in a reducing

atmosphere for a single - component metal powder compact can­

not be described in a simple manner as it depends on the

surface condition of the powder. There is a considerable

variation in activation energy which determines the rate of

 sintering, and this depends strongly on the sintering

atmosphere. However, it may also he stated that, (a) sinter­

ing in vacuum results in minimum porosity, (b) the vacuum

sintered compacts have the minimum number of fine, and a

maximum number of coarse pores, whereas in the hydrogen

sintered compacts the maximum number of pores lies in the

medium pore size range, (e) it is also known that the pore

size distribution determines the grain size in the sintered

compact, and it may therefore be concluded that the sintering

r
atmosphere directly affects the final grain size of the sintered

pu
ar t
compact.
h
However this theory is seemingly applicable for single
ag
Kh ig
phase components. What happens when sintering proceeds through

a number of intermetallic phases and where a transient liquid

r

phase appears and disappears as in case of sintering of

IIT py

90 copper / 10 tin components is not established. The work of

68
Fleming , however does give a general idea as to what effect

should be obtained if the sintering of 90 copper / 10 tin

o

bronze bearings are to be carried out under a vacuum atmosphere,

C

instead of the conventional reducing gas atmospheres.

AIM AND SCOPE OF THE PRESENT WORK

With the present global oil and natural gas crisis,

alternative to conventional processes (using petroleum or its

by products) are being hotly persued in every fie ld . A

glance at the warnings that have been issued by various bodies

 34
to conserve petroleum and its by products, drives the point

home that more stringent measures are to be taken for

conserving the limited resources. One of the reports published

in the " World Energy Resources 1985-2000" dealing with oil

69
and gas resources says that, H By raising the problem of

o i l supply to the industrialized countries, the events of

recent years have drawn attention to the problem of the

exhaustion of worldwide petroleum resources. Until then the

r
public had been only vaguely aware that whereas mankind had

pu
only about 30 years of petroleum reserves, a short and
ar t
alarming length of time, these thirty years were constantly
h
moving forward, ..........
ag
Kh ig
These worries have led to a proliferation of

futuristic surveys during the last two or three years. The

r

opinion that is generally expressed is that by the end of

IIT py

the century the world is in serious danger of lacking crude

o il as the result of a decrease in production that should

o

begin around 1990. The most pessimistic, combining the

p o litical dangers of a slowing down of production and the

C

assumption of relatively high economic growth rate, even

consider that difficulties of supply w ill appear during

the coming decaden .

The recent Iraq Iran war has added a new dimension

to this grave problem and has also resulted in a spurt of

price rise of crude petroleum barrels.

 So, to reduce the dependence on petroleum products

or natural gases which are obviously becoming dearer (and its

resources are also dwindling) it was decided to try vacuum as

the sintering atmosphere for one of the most conventional

powder metallurgy product, namely the self lubrication bronze

bearings.

In recent years vacuum technology is taking large

strides and is gradually encroaching into totally alien

territories. Specially with the advent of continous vacuum

r
furnaces in the market, the main argument which was used

pu
against the vacuum sintering atmosphere,
ar t 11 that its a batch
h
process" , is fast loosing its ground. The interaction of

market needs and equipment development is always giving new

ag
Kh ig
solutions to old problems and creating openings to tackle

completely new problems. This has led to a rapid development of

r

vacuum furnaces and their applications. It is now in the offing

IIT py

that a number of P/M part manufacturers will using vacuum

as the sintering atmosphere in the near future.

o

Vacuum furnaces have already entered the metallurgical

C

market in a big way owing to its obvious advantages. The boom

in the aerospace industry has called for an expension in the

vacuum heat - treatment furnaces. It is on the cards that we

shall see further advances in vacuum heat-treatment technology

in the near future. Already liquid quenching vacuum furnaces

70
are coming to the market as described by Buckley . In an
71
article in Metallurgia it is stated that " the whole
 36

industry is in a state of technological evolution in which

vacuum heat-treatment w ill doubtless play an increasingly

important r o l e ” . A new vacuum furnace is also reported in

72
M etallurgia which is a twin zone one, believed by its

manufacturers, Wentgate Engineers, to be the fir s t of its kin d.

Vacuum sintering furnaces are also making inroads

73
into the powder metallurgy industry . A new model of vacuum

furnace, it is claimed, provides more uniform sintering

r
temperature. Batch vacuum units from GCA International (Euro­

pu
pean representatives of GCA Vacuum Industries, USA) is
ar th
introduced to fa c ilit a te complete cycling processing of

sintered hard metals. The new furnace offers + 5°C temperature

ag
Kh ig
uniformity throughout its 36 inch long, 18 inch wide, 12 inch

high work zone.

r
IIT py

Another continous vacuum furnace for powder metallur­

gica l products is designed by Degussa. This furnace consists

o f three zones , one pre-sintering zone under atmosphere,

o

one sintering zone and one cooling zone under vacuum. In an

74
C

artic le by Bauer on recent developments in automated

vacuum sintering, he says, 11 During the last few years,

Degussa has presented the powder metallurgy world with another

greater vacuum sintering furnace for hard-metals which is

o f double the capacity compared to the former one. The

automated vacuum sintering plant, called 'Sinterautomat 1

is a very good solution for sintering mass produced hard metal

parts M .
 37

The advantages of vacuum atmosphere over the other

normal gas atmospheres are numerous

(1 ) Excellent temperature uniformity,

(2 ) No environmental pollution by toxic m aterials, heat

or fumes,

(3 ) Less distortion problems,

(4 ) Easier material degassing and complete elemination of

surface oxidation or deterioration,

(5 ) Significant reduction in labour costs both in the

r
furnace operation and by elemination of subsequent

pu
cleaning operation,
ar th
(6) It gives much more e ffic ie n t energy usage.

ag 75
Kh ig
Tamalet in his article on " How to decrease the

Energy Demand of Powder Metallurgy Furnaces” has s a id that

r

11 we cannot refer to atmosphere without mentioning vacuum.

IIT py

This solution is obviously ra d ic a l,t o cut 50% of energy

costs in the sintering operation" , He has described a

vacuum sintering furnace the ’Metafram Sedimec Furnace*,

o

which is a continous car type furnace. It has a pre-heat zone,

C

a process zone, a cooling zone and two locks one for product

entry and the other for e x it . The furnace production may

o
reach 150 kg/hour and a maximum temperature of 1400 C. The

products obtained in this furnace always shows better

properties than those obtained in conventional furnaces.

 38

How the heat-transfer in vacuum takes place is

76
discussed in details by Limque et al . According to them

vacuum plays a double role in terms of heat tran sfer.

Comparing vacuum atmosphere with gas atmospheres (eg. Argon)

it is said that in vacuum the heat insulation efficie ncy of

ceramic wool is increased by 6 0 $ . Additionally, better

insulation and lower heat losses allow lower heat input in,

a state of equilibrium, with consequent smaller temperature

differences. The temperature distribution in a loaded

r
o o

pu
chamber is maximum + 5 K even at 600 C.
ar th
From the following discussions it is obvious that

vacuum technology has a lot to offer to the powder metallur­

ag
Kh ig
gists in form of sintering atmospheres. The shift to vacuum

atmosphere is further prompted by the energy conservation

r

drive, low petroleum and natural gas resources and soaring

IIT py

price of petroleum and its byproducts. Obviously, from the

energy and product quality viewpoint, the vacuum furnace

gives, by far the best results, but its use is as yet

o

lim ited owing to its price being much higher than the
C

conventional furnaces. But, as more and more manufacturers

s h ift their alliance towards vacuum atmosphere, it is

expected that the price of the furnaces w il l f a l l rap idly.

One of the good signs for vacuum technologists is the number

o f big names in the furnace manufacturing world laying

stre ss, and committing themselves to huge f is c a l investments

for the development of vacuum furnaces.

 39
So, with the developments in the continous

vacuum sintering furnaces, and with the ever increasing

cost of production by the conventional route, it becomes

timely to investigate the fe a s ib ilit ie s of production of

porous bronze bearings through vacuum sintering.

It is proposed to look into the fe a s ib ilit y of

sintering se lf lubricating bronze bearings in vacuum. The

investigations would cover the influence of high and low

vacuum on properties of bearings produced and also help in

r
determining the type of vacuum set up essential for proper

pu
sintering of such m aterials. The other important investiga­
ar th
tions w ill include evaluation of types of copper powders,

the suitable pre-sintering and sintering temperature and

ag
Kh ig
time, the heating rate, the compacting pressure, the tin

contents that would be suitable for vacuum sin terin g . It is

r

also proposed to compare some of the vacuum sintered bronze

IIT py

bearings with bearings sintered under similar conditions but

in hydrogen atmosphere.
o

Some of the properties to be determined are the total,

C

open and closed porosities as a percent of volume ; closed

porosity as a percent of total porosity ; sin tered density ;

percent theoretical density and the K-strength constants. It

is also proposed to look into the structures to try and

understand its effects on the fin a l properties of porous

bronze bearings. This work may help to establish vacuum

sinterin g as an alternative to the conventional gas sinter­

ing of self- lubricating bronze bearings.