PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.

DATE DUE DATE DUE DATE DUE

‘
II

 

 

 

 

 

 

 

—_T

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

=é—lg

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

__l[_______

 

 

 

 

MSU Is An Afflrmdive Action/Equal Opportunity Inuitution
ottoman”:

A.HETALLIC-GLASS RIBBON REINFOREED GLASS-CERAHIC

MATRIX COMPOSITE

BY

Rajendra Uddhav vaidya

A THESIS

Submitted to

Michigan State University

in partial fulfillment of the requirements

for the degree of

DOCTOR OF PHILOSOPHY

Department of Metallurgy, Mechanics and Materials Science

1991

A.METALLIC-GLASS RIBBON REINFORCED GLASS-CERAMIC

MATRIX COMPOSITE

BY

Rajendra Uddhav vaidya

The feasibility of producing metallic-glass
ribbon reinforced glass-ceramic composites by low cost
pressing and sintering techniques was demonstrated.
The quality of the composite specimens produced was
affected by factors such as binder content, compaction
pressure and sintering temperature. The mechanical
behavior of such composites was studied and the failure
modes could be correlated to variations in the matrix
strength and interfacial bond strength with
temperature.

The fracture toughness of such composites was
dependant on the ribbon orientation (with respect to
its long and short transverse faces) and ribbon layout
in the matrix. This was primarily attributed to

differences in the moments of inertia of the ribbons in

the different orientations. The load transfer
characteristics of this composite system was found to
be different from those of other fiber reinforced
systems. The ribbon width was found to play a
significant role in the stress transfer between the
component phases.

Discontinuous metallic-glass reinforced glass-
ceramic matrix composites were also fabricated. The
mechanical strength of such composites was found to be
within 80% of the strength of a continuously reinforced
composite containing the same volume fraction of
reinforcements. The strength of such composites was

also found to be isotropic in the plane of compaction.

First and foremost I would like to thank my
adviser Professor K. N. Subramanian for all his
encouragement, help and support, without which this
work would not have been possible. I would also like
to thank my comittee members Professor E. D. Case and
Professor N. J. Altiero of the Department of
Metallurgy, Mechanics and Materials Science and
Professor H. Eick of the Department of Chemistry for
their valuable suggestions from time to time. I would
also like to express my gratitude to Dr. L. T. Drzal
and the Composite Materials and Structures Center at
Michigan State University for supporting and funding
this project. Thanks are also due to Professor K.
Mukherjee and all other faculty, staff and students of
the Department of Metallurgy, Mechanics and Materials
Science for their support. Last but not least, I would

like to thank my parents and sister for their support.

CHAPTER 1:

CHAPTER 2:

CHAPTER 3:

CHAPTER 4:

CHAPTER 5:

APPENDIX:

.IABLB_QE_QQNIENI§

Fabrication of metallic-glass ribbon
reinforced glass-ceramic matrix

composites.................................1

Elevated temperature mechanical properties
of metallic-glass ribbon reinforced

glass-ceramic matrix composites... ..... ...38

Effect of ribbon orientation on the fracture
toughness of metallic-glass ribbon
reinforced glass-ceramic matrix

comPOSitBSOOOOOOOOOOOOOCOCO...00.000.000.058

Load transfer mechanism in metallic-glass
ribbon reinforced glass-ceramic matrix

comPOSiteSOOOOOOOOOOO00.0.0000...00....00.79

Discontinuous metallic-glass ribbon
reinforced glass-ceramic matrix

comPOSiteSOOOOOOOOOOOOOOOOOOOOO0000......105
Mechanical properties of metallic-glass

ribbon reinforced glass-ceramic matrix

comPOSiteSOOOOOOOOOOOO00.0.00000000000000131

IV

Metallic-glass ribbon reinforced glass-ceramic
matrix composites were fabricated by conventional wet
pressing and sintering techniques. Some of the
important aspects of composite fabrication have been
discussed here. Variables such as binder content,
compaction pressure and specimen size, and their
interdependancy on one another are discussed. The
optimum processing conditions required to obtain the
best specimens will be stated. The sintering
temperature was also found to be very important. The
sintering temperature not only controls the softening
characteristics of the parent glass, but also affects
the specimen distortiona Proper selection of the
sintering temperature can ensure minimum distortion in
the specimens, while at the same time retain the ease

and efficiency of fabrication.

IEIBQDQQIIQE=

Developments in fiber reinforced ceramic matrix
composites have proceeded at a rapid pace in the past
few years [1]. Various types of fibers, especially
silicon carbide, alumina and carbon, have become very
popular as reinforcements for these brittle ceramic
matrices because of the high strength and toughness
. they impart when incorporated into the ceramic

matrices.

The three classes of ceramic matrices commonly
employed are glasses, glass-ceramics and crystalline
ceramics. Glass matrices, although easy to fabricate
because of the softening they exhibit, are also
handicapped by low useful operating temperatures for
the very same reasons. Crystalline ceramic matrices
are relatively difficult to process, but have
relatively higher use temperatures. These matrices are
usually fabricated by powder pressing techniques which
invariably result in some amount of retainedporosity in
the structure. This porosity although small in
magnitude, can result in a severe deterioration of the
mechanical properties of the system. Techniques such as

hot-isostatic pressing, vapor deposition and sol-gel

method have been employed to either obtain a finer

particle size or densify the structure, but have only
been achieved at the expense of increased costs and/or
low productivity rates (because of the relatively long

processing times).

Glass-ceramics combine the advantages of both
the glassy and crystalline materials. The glass-
ceramics are fabricated in the glassy state, and hence
relatively lower temperatures can be used as compared
to those required in the fabrication of the crystalline
ceramics. Upon fabrication, the glassy matrix is
converted to a crystalline state. This crystalline
state does not exhibit softening, and hence the
resultant glass-ceramic can be used at temperatures
close to the melting point of the crystalline phase,
which is usually much higher than the softening point
of the parent glass. Proper selection of the parent
glass can result in a glass-ceramic during the
fabrication step itself, saving time and cost for a
separate heat treatment step. These glass-ceramics also
have a significantly lower amount of porosity as
compared to crystalline ceramics processed under
similar conditions. Various glass-ceramics being
actively investigated recently include LAS (Lithium

alumino silicates) and MAS (Magnesium alumino silicate)

systems.

Metallic-glasses are an interesting class of
materials being studied. Metallic-glasses are produced
by rapid solidification of molten metal. High rates of

cooling are employed for the same (of the order of 106

to 109 °

K/sec), and crystallization of the molten metal
is suppressed. The resultant structure of the material
has no long range order and is referred to as being

‘amorphous'.

Metallic-glasses possess unique mechanical
properties. They are extremely strong and hard, and are
very tough. They are also more oxidation and corrosion
resistant as compared to the parent metal from which
they are derived. The ribbon geometry is a natural
consequence of the high cooling rates required in the
manufacture of these metallic-glasses. This ribbon
geometry provides for an extremely large surface area.
Another interesting feature of these metallic-glasses
is their ability to be bent to almost a zero radius
with very little plastic deformation. This ability of
the metallic-glasses makes them extremely damage

tolerant.

Metallic glasses have been investigated as

reinforcements for polymer [2-5] and metal [6]
matrices. These studies indicated that significant
improvements in the mechanical properties could be
achieved with‘relatively small volume fractions of
reinforcements. The potential of using metallic-glasses
as reinforcements for brittle glass-ceramics was not
demonstrated until recently [7-12]. These studies also
indicated the tremendous potential of developing such
composites for structural applications. In addition to
the improved properties, it should also be noted that
relatively easy and cost efficient techniques were used
in the fabrication of the same. Both continuously and

discontinuously reinforced composites were fabricated.

In this paper, some of the aspects of
fabrication of such metallic-glass ribbon reinforced
glass-ceramic matrix composites will be presented. The
various fabrication parameters and their effect on the
fabrication processes will be discussed. Some of the
practical problems encountered in the fabrication of
such composites will also be emphasized. The important
point to keep in mind here is that all of these
experiments resulted in the fabrication of laboratory
sized small specimens. No commercial component was
fabricated. Hence problems related to scaling up will

have to be considered seperately if such composites

have to be fabricated for commercial applications.

Two commercially available metallic-glasses
namely METGLAS MBF-75 and METGLAS MT 26058-2 were
selected as the reinforcements. This selection was
based on two important factors; good mechanical
properties and relatively higher crystallization
temperatures. METGLAS is a trademark of Metglas Corp.,
a subsidiary of Allied Signal Ltd for brazing metals
and alloys, and both the metallic-glasses used were
obtained from them. Since the production of METGLAS
MBF-75 has been discontinued, most of the latter
studies were carried out on composites incorporating
METGLAS 26058-2. The physical and chemical properties
of these metallic-glasses as provided by the

manufacturer are given in Table l.

The crystallization temperature of the metallic-
glasses dictated the choice of the selected glass
matrices. Both the metallic-glasses had

crystallization temperatures in the range of 550°C.

TABLE 1: Physical properties and chemical compositions

of the metallic-glass ribbon reinforcements.

 

 

Property Metglas 26058-2 Metglas MBF-75
Chemical Fe 78 Ni 50
composition (t) B 13 Co 23
81 9 CrlO
Mo 7
Fe 5
B 5
Crystallization
temperature(°C) 550 605
Elastic modulus 85 70
(GPa)
Yield strength >700 1300
(MPa)
Coefficient of 76 x 10'7 78 x 10’7
(°c‘1)
Density (g cm-1) 7.18 7.46

 

Hence it was critical to select glass matrices having
'softening temperatures below 550°C. Heating up a
metallic-glass above its crystallization temperature
usually results in a loss of mechanical properties. In
fact it has been shown that the mechanical properties
of this metallic-glass begin to degrade far below its
crysatllization temperature (Table 2). However the
loss of strength was not that significant, and a large
portion of the initial strength and toughness of the
metallic-glass can be retained if the temperature is
maintained loo-150°C below the crystallization

temperature of the metallic-glass.

Based on all of these requirements, Corning
Glass Codes 7572 and 8463 were initially selected as
potential matrices. However, based on mechanical
property considerations, the 7572 matrix was found to
be superior as compared to the 8463 matrix. Subsequent
tests were conducted on specimens fabricated using the
7572 matrix only. Some of the properties of the
matrices as provided by the manufacturer are given in

Table 3.
Ma

The metallic-glass reinforced glass-ceramic

10

TABLE 2: Measured tensile strength (in MPa) of the

METGLAS 26058-2 ribbon as a function of

temperature and time.

 

 

time (min)
0 20 30 60 120 180 300
ten“ C)
100 -- 2059 1464 2236 --- 1426
200 -- 1987 1842 2114 --- 1420
300 -- 1894 1733 2328 --- 1905
400 -- 1028 1102 986 972 ---
450 1709 -- 775 851 724 ---

550 -- 443 -- --- -—- ---

 

H

TABLE 3: Physical properties and chemical compositions

of the glass matrices as provided by the

 

 

manufacturer.

Property Corning Glass 7572 Corning Glass 8463

Softening point 375 370

(°c>

Coefficient of 95 x 10'7 105 x 10’7c

thermal expansion

(°c'1)

Density (powder) 3.8 3.8

gm/cc (fired) 6.0 6.2

Continuous service 450 450

temperature (°C)

Chemical composition Pbo 70 Pbo 84

(t) B 03 5-10 B 03 5-10
830 2-5 sic 2-5
Al 3 1-5 Al 5 1-5

2n3 310-20 zn5 310-20

 

12

matrix composites were fabricated in steel dies. Two
different geometries (cylindrical and prismatic) were
employed for the purpose. In the prismatic geometry two
different sizes were used.The steel dies were
fabricated out of medium carbon alloy steel (4340), and
heat treated to a hardness of 50 Rc' The dies with the
rectangular cavities had a ‘split' arrangement in order

to facilitate specimen removal (Figure 1 a and b).

A Tinius-Olsen machine was used for compacting
the powders in the dies. The machine had a maximum
rated capacity of 60,000 lbs. Figure 2 shows the

experimental setup used.

Wiggles:
1!) .WLW=

The metallic-glasses were obtained in the form
of long continuous ribbons, about 2 inches wide (Figure
3). The ribbon was wound on a spool. The average
thickness of the ribbons as measured was 25 microns.
The ribbons were cut using an extremely sharp pair of
scissors. One interesting point to note was that

cutting the ribbons in a direction perpendicular to its

 

FIGURE 1: Dies used in the fabrication of the composite
specimens.
a. Different sizes and geometries used

b. Split die assembly

 

FIGURE 2: Experimental setup used in composite specimen

fabrication.

 

FIGURE 3: The metallic-glass ribbon reinforcement

spools as obtained from the manufacturer.

16

width resulted in a relatively jagged and irregular
cut. 0n the other hand, cutting the ribbons along the
width was relatively easy, and did not lead to an

irregular edge.

Cutting the ribbons caused them to bend slightly
towards their inside (side in contact with the rollers
during fabrication). This bending was an important
factor which had to be accounted for during specimen
fabrication. All the ribbons were thoroughly cleaned

with acetone before being incorporated into the matrix.

Amyl acetate was used as the binder for the
glass powder. This binder was recommended by the
manufacturer (Corning Inc.). However the amount of
binder used varied from the amount suggested by the
manufacturer (3% by weight). Various factors
influencing the amount of binder used included the size
of the specimen and volume fraction of ribbons being

incorporated into the specimens.

Care was taken to ensure complete binder removal
before sintering the specimen. The specimens were
preheated to 250°C for sufficient periods of time
before being sintered. In cases where larger amounts of

binder were used, even longer times were employed.

l7

Binder removal was determined relatively easily because
of the extremely strong odor which was emitted during
vaporization of the binder.

In order to fabricate the continuously
reinforced composites, the glass powder was weighed
out, and mixed with appropriate amounts of amyl acetate
binder. A small amount of powder was placed in the die
and pre-pressed. The sized metallic-glass ribbons were
then laid down within the die over this pre-pressed
lower layer. Care was taken to see that the concave
surface of the ribbon (bent) always faced upwards.
Avoiding this led to the formation of cracks in the
green compacts. After the required volume fraction of
ribbons were incorporated into the matrix, a final
layer of matrix material was placed at the top and the

whole compact was pressed.

The compaction pressure was applied very slowly
in order to ensure uniform density distribution within
the compact. The amount of pressure used depended on
two important factors; the thickness of the specimen

and the binder content.

After compaction, the specimens were removed

from the die. The ‘split' assembly of the dies made

18

specimen withdrawl extremely easy. However, sometimes
the specimens did stick to the die cavity. Use of a
light machine oil to lubricate the die cavity was found
to remedy the situation, although a small amount of oil

did get absorbed within the specimen.

The green compacts were then laid out on
stainless steel or alumina plates. In order to prevent
the specimens from sticking to the plates, they were
either coated with a thin film of graphite or covered
with an extremely thin foil of aluminum, which was

easily peeled off after fabrication (Figure 4).

As mentioned earlier, sintering of the specimen
was preceeded by a binder burnoff stage. Sintering was
carried out in an air atmosphere at 400°C. The
sintering time varied depending on the specimen, and
usually ranged between 1-5 hours. The sintering stage
was accompanied by a heat treatment at 450°C, which
converted the glassy matrix into a crystalline state.
Time for the crystallization treatment varied between
20 minutes and 1 hour. The main crystalline phase was
2Pb0.Zn0. After crystallization, the specimens were

furnace cooled down to room temperature.

All the specimens experienced an adequate amount

 

FIGURE 4: Loading the green compacts in the furnace for

sintering.

20

of flow during the sintering stage. Due to this all the
specimens had rounded upper edges. Some amount of
polishing was required to restore the required shape.
However, the dimensions of the polished specimens
were always within 30% of the dimensions of the

corresponding green compacts.

b) .W:

The procedure used in the fabrication of these
composites was more or less similar to the one
described earlier. The metallic-glass ribbons were
sized to appropriate dimensions. After sizing, the
ribbons were precoated with the glass powder.
Precoating was done by preparing a slurry of the glass
powder with excess of amyl acetate binder, and
incorporating the ribbons within the slurry. The slurry
was then heated up to 250°C to remove the binder. This
was then mixed with more glass powder containing a
relatively smaller amount of binder. After mixing,
compaction in the die was carried out, followed by
sintering and crystallization. In spite of the higher
pressure used, the green compact strength of the
specimen was very poor and more than 60% of the
specimen had to be discarded because of cracking

problems.

.DIEEEEEIQN3

The variation in the compaction pressure with
binder content is shown in Figure 5. From the figure it
can be seen that the compaction pressure decreases with
increasing binder content. However for pressures
exceeding 5000 psi, pressure cracking was observed in
the specimens, irrespective of the binder content used.
Compaction pressures less than 200 psi were also found
to be insufficient to impart enough green strength to
the compacts. Binder contents in excess of those
determined by the curve led to the excess binder being
squeezed out of the sides of the die. The combination
of binder content and compaction pressure used are
indicated in the figure.

The variation in the time required for binder
removal with preheat temperature is shown in Figure 6.
Shorter preheating times were required at higher
temperatures. However, the preheat temperature had to
be maintained below the softening temperature of the
glass matrix in order to prevent any pores from getting
trapped in the matrix during the glass flow. A
temperature of 250°C was found to be most suitable.
Higher temperatures led to shorter preheating times,

but also caused cracks to form in the compacts,

21

22

  
       

 

  

binder

 

 

 

 

squeeze out ._1,
.3
a +3
.. o
E 3 a
'3 inadequate 2
Q green strength a
3 2» f
a.
l-’

 

 

 

i i 3 4 5
compaction pressure

(x 103p”)

FIGURE 5: Variation in the compaction pressure with

binder content.

temperature (C)

23

j,

400"

xcessively fast
specimen crack

 

 

a
o
o
L
k

   

 

excessively slow 1
- long times

100 -

 

 

3'0 6'0

time(min)

 

FIGURE 6: Variation in the time required for binder
removal as a function of the preheat

temperature.

24

probably as a result of the binder vaporizing too

quickly.

Heating the compacts to the softening point and
above led to glass flow, which in turn caused the
specimens to change in dimensions. Flow of the glass
matrix usually caused the specimen thickness to
decrease, and specimen width to increase. This flow in
the specimens was measured by calculating the change in
the specimen thickness as compared to the thickness of
the original green compacts. The specimen dilation as a
function of increasing time, for different temperatures
is shown in Figure 7. For temperatures below the
crystallization temperature of the glass matrix,
increasing times led to increased glass flow. At the
crystallization temperature, the specimen dilation
increased initially, but then decreased as the
crystallization process progressed. Crystallization of
the matrix led to an increase in the viscosity of the

matrix.

The equation one can use to predict the
dilation due to matrix flow (between 370°C and 450°C)

has the form

25

 

 

 

 

30" o
’5
2 I.
X
zI-°
C o
.2
3’. ' o
'5
‘0‘ 370 c
5‘
1&0 150

FIGURE 7: Specimen distortion as a function of time

for different temperatures.

26

R = c exp [31: - D/t].............(1)

where, R is the specimen dilation defined as
(change in specimen height)/(original height)
t is the time, and

B, C, and D are constants based on curve fitting

For t = 0, the dilation is zero, and

for t

infinity , the dilation is infinite.

For very large times, the equation can be approximated to

R = Cexp [BtJOOOOOOOOOOOOOOOOOOOO(2)
For a given temperature and large times, B and C
can be taken as constants. In order to determine these

constants differenciate both sides of equation 2.

Hence, d(R)

 

B C exp (Bt)
d(t)

27

Hence, d(R)

- B R = 0.....................(3)

 

d(t)

This equation can be used to determine the
constants B and C for different temperatures. Choosing
points in the regions corresponding to longer time
periods in each curve in Figure 7, the values of B and

C can be obtained as

1

Temperature B (sec- ) C
370°C 7.14 x 10'3 16.52
400°C 0.01 9.197
420°C 0.017 2.314

To obtain the value of D, one can use Equation 1
and choose points in the regions corresponding to
shorter times of the curve (assuming B and C remain

constant). The values of D obtained are

28

Temperature D (sec)
370°C 87.30
400°C 36.15
420°C 0.81

For temperatures below the crystallization
temperature, one can assume that the viscosity of the
matrix is independant of time. However at 450°C, the
viscosity becomes a function of time and temperature.
This is because as time progresses, more and more of
the matrix crystallizes and causes the viscocity to
increase. The equatioanor the dilation takes the

form

R = C(T,t) exp [B(T,t)t]...........(4)

As the temperature increases the pre exponent
term C decreases and its effect on the dilation becomes
less important as compared to the exponent term. Hence

we can assume that C is a constant (k) for practical

29

purposes. Equation 4 then takes a form

R = k exp [B(T,t)t]....... ..... .....(5)

Determining B here is difficult because it is a
function of time and temperature. In the limiting case

for two times t and t2 very close to one another

1

 

R1 exp [Btll
— = ~......... ............. (6)
R2 exp [BtZ]

Equation 6 can be used to find out the value of
B as a function of temperature. This can be carried out
by choosing points in various portions of the curve
very close to one another. Variation in the constant B
with increasing time is shown in Figure 8. From the
figure it can be seen that B decreases as a function of

time as a result of the crystallization process.

These results are useful for determining the

3D

 

 

0.1 "’
In 0.05 "'
0.025 "'
10 2'0 3'0 4'0
time (min)

FIGURE 8: Variation in B as a function of time at

450°C.

31

optimum sintering temperature and time. For sintering
it is necessary to have adequate flow but not too much
so as to cause extensive distortion in the specimens.
Sintering at the softening temperature of 370°C was
found to take too long. 0n the other hand sintering at
450°C was completed in a short period of time but was
accompanied by a large amount of distortion.
Crystallization of the glass matrix occuring at this
temperature was also found to be detrimental to the
fabrication process (because crystallization was
accompanied by an increase in the viscosity). Specimens
fabricated at these temperatures were found to exhibit
features similar to "cold shuts" (Figure 9).
Fabricating at 400-420°C was found to give the best

specimens (Figure 10).

Some of the features observed on the fracture
surfaces can be seen in the accompaning scanning
electron micrographs (Figures 11 a, b, c and d).
Important features to note in these micrographs are the
strong bonding between the matrix and the ribbon

reinforcements, and negligible porosity in the matrix.

32

 

...l

 

.. «Lia-Try almanwvuflwi ' ~. ~ A e 9“.“ J
'h ..5‘ .: E ",-u’(i- ‘ . .3 . _I~_
I if , i ‘ 7 ° :
1 g ‘
a at mww fl flip-"f

3’1 ism—-

 

—1. F'Fa'cw '_
may; ‘3. ‘

.53

450

FIGURE 9: Inadequate glass flow in specimens sintered

at 450°C.

33

use 31:92:11.2} 0.1.13.3r 3

w

gar. ..,3‘iflr.§5§33.i9.3
.25.. ,« ... , .. .. ,
m .— . ,

3.

'2‘:

.....
...
. , , Z.
, . , at”?
.. , . . , ... .... . .c. "L... 2.2,.5. 2 2 z .. ... .... .mcx;
.42.eEEeéh..fig hascmuhflfirmfilierri 41.33%

u . :,..1.E&......:..~e_.... as... 431...... .,

I. . «Erasmus...» .. . I. ..v 2.. . .

.. garignfi I. n. w! . ... . z . 2.. . ,.

. firm?” .3412: , , .2 .... a. ... 5%..." {do r. . ,: ..u e ; ,
. ___,.. .e...c,,.~. u. . . ,

  

. . . . _
:1 . . .. . u
r _ . .
w . , _, .. 1 _

. .. 2 .
. w: .2 . . . .. .. 1x. . M .. ...r r ... f 2 2.1-, ....In ..... .1
.... ,. . ......f vi, ... Libya; 1.5.7.... Arr)... “was“:“fq. a 11..."..92xwmsl ..,".. 2mm”. «1. .1.
3.5.1.51. stinger”... A...

1K?

400

Specimens sintered at 400°C.

FIGURE 10

34

 

 

FIGURE 11: Fractured surfaces of the composite specimens.
a. Presence of good ribbon-matrix bonding.
b. Matrix material adhering to the ribbon surface.
c. Good ribbon-matrix bonding and negligible
porosity in the matrix.

d. Multiple ribbon composite.

This paper dealt with some of the aspects of the
fabrication of metallic-glass ribbon reinforced glass-
ceramic matrix composites. It has been shown that such
composites can be produced by relatively low cost wet
pressing and sintering techniques. Variables such as
sintering temperature, binder content and sintering
time are important and need to be carefully controlled

to obtain good specimens.

The authors would like to acknowledge the
support of the Composites Materials and Structures
Center at Michigan State University. The authors would
also like to thank Mr. Micheal Brykalski for the data

provided in Table 2.

36

10.

11.

RE CES:

Tai-il Mah, Mendiratta M., Katz A. and Mazdiyasni 8.,
Ceramic Bulletin, Vol.66, No.2, (1987) p. 304.

Strife J. and Prewo K., AFWAL Report TR-80-4060, 1980.
Strife J. and Prewo K., Journal of Materials Science,
Vol.17, (1982) p. 359.

Fels A., Friedrich K. and Hornbogen E., Journal of
Materials Science Letters, Vol.3, (1984) p. 569.
Friedrich K., Fels A. and Hornbogen E., Composites
Science and Technology, Vol.23, (1985) p. 79.

Cytron 8., Journal of Materials Science Letters, Vol.1,
(1982) p. 211.

Vaidya R. U. and Subramanian K. N., Journal of Materials
Science, Vol.25, (1990) p. 3291.

Vaidya R. U. and Subramanian K. N., ASM/ACCE/ESD
Conference Proceedings, Detroit, 0ct.89 published by ASM,
Metals Park, Ohio, p. 227.

Vaidya R. U. and Subramanian K. N., Journal of Materials
Science (in press).

Vaidya R. U. and Subramanian K. N., Journal of American
Ceramic Society (in press).

Vaidya R. U., Lee T. K. and Subramanian K. N., ASC

Conference Proceedings, E-Lansing, June 90, published by

37

38

Technomic press, Lancaster PA, p. 902.
12. Vaidya R. U. and Subramanian K. N., Composites Science and

Technology (in press).

CHAPTERZ

39

Effect of temperature on the modulus of rupture
of glass-ceramic matrix composites reinforced with very
small volume fractions (”1%) of continuous metallic-
glass ribbons was studied. The failure modes in such
composites were found to significantly depend on the
test temperature. Variations in the strength of the
matrix, and the interfacial shear strength between the
matrix and the ribbons due to changes in the test
temperature, were found to control the elevated
temperature strength and failure mode of these

composites.

40

IEIEQDQQIIQN=

Reinforcement of brittle matrices with high
strength fibers results in dramatically improved
mechanical properties such as increased toughness and
high strain to failure. Recent developements [1-6]
have led to a renewed interest in continuous fiber
reinforced ceramic matrix composites for high
temperature applications. Metallic-glass
reinforcements in the form of ribbons possessing large
surface areas and large aspect ratios have been
demonstrated to be very effective in enhancing the
mechanical properties of glass-ceramic matrices [7,8].
Even very small volume fraction of metallic-glass
reinforcements were found to improve the strength,
elastic properties and fracture toughness of the glass-
ceramic matrices quite significantly. One of the
potential use of such composites is at elevated
temperatures, since glass-ceramics do not exhibit
softening, unlike the parent glass from which they are
derived. Successfull application of such metallic-
glass ribbon reinforced glass-ceramic composites for
high temperature applications depends on the
characterization of their mechanical response at

elevated temperatures. An understanding of the failure

41

42

made at elevated temperatures is also important to make

"life-time” predictions.

 

Rectangular bar shaped composite specimens (5 cm
x 1.5 cm x 0.4 cm) were prepared by using Corning Glass
Code 7572 as the matrix and METGLAS MT 26058-2
continuous ribbons as the reinforcements. The specimens
were prepared by cold pressing and sintering
techniques. Details of the chemical compositions of
the ribbon and matrix, and composite manufacturing

techniques have been described elsewhere [7].

Three point bending tests were carried out in an
Instron testing machine with a cross head speed of 0.05
cm/min, in accordance E6 A.8.T.M standard C-203/85.
Elevated temperature tests were carried out by heating
the specimen and the test fixture, with an electrical
resistance furnace. The temperature of the furnace was
controlled to within + 5° C. The specimens were soaked
at the test temperature for 30 minutes prior to the
tests. All the bending tests were performed in a
nitrogen atmosphere to prevent oxidation of the
fixture. A total of three specimens were tested at
each temperature. Pullout tests were also carried out
at various temperatures to measure the ribbon-matrix

interfacial bond strength.

BESEHI§_AND_DI§£!§§IQE=

Plots of the Modulus of Rupture (MOR) of the
unreinforced matrix and composite specimens as a
function of test temperature are provided in Figure 1.
All the specimens were observed to retain most of their
room temperature strength even at about 400°C. The
composite specimens exhibited higher MOR values as
compared to the unreinforced glass-ceramic matrix
specimens at all temperatures tested. At higher

temperatures matrix creep became predominant.

The load-deflection curves of the unreinforced
matrix and composite specimens (containing 1.25 volume
percent of reinforcements) at various temperatures are
provided in Figures 2 and 3 respectively. At lower
temperatures (upto 200°C) the composite exhibited
typical brittle behaviour, with no evidence of ribbon-
matrix debonding and sliding. At 300°C, a dramatic
change was observed in the load-displacement curve,
with the composite exhibiting ribbon debonding and
sliding, accompanied by a large elongation to failure.
At 400°C, although the matrix began to creep, the
composite still retained a large portion of its

strength. At 500°C, extensive matrix creep occured and

43

44

 

 

 

 

1

£35“ 1-

1:

v30- 1

o

§E25 1‘

g i

“20- L y

B t

“15"

3.

:10"

o

2 5-
0 l I l l i
0 100 200 300 400 . 500

Temperature(C)

FIGURE 1: Plot of the Modulus of Rupture of the matrix
and composite specimens as a function of
temperature.

( 0 Matrix, A Composite with 0.73% by volume of
ribbons and II Composite with 1.25% by volume of

ribbons.)

45

 

 

 

 

50—
RT.

40- 1 300° 0
530- 400° 0
‘8
3 20-

' 0
W 500 0;
0 1'0 1‘5 2‘0 2'5 3'0
Displacement
(10"3 cm)

FIGURE 2: The load-displacement curves of the matrix

specimens at various temperatures.

" 46

 

 

 

 

R.T.
100i \ 300°C
‘ / 400°C
"2‘75-
13 ,
gso— . L i
" 25_ 500° 0 *

 

 

i

1 1 l I I I I I F
20 40 60 80 100 120 140
Displacement

(IO—3cm)

FIGURE 3: The load-displacement curves of the

composite specimens (Vf=1.25%) at various

temperatures.

47

the composite failed at a very low load.

The observed changes in the load-displacement
curves is an indication of the change (decrease) in the
ribbon-matrix interfacial bond strength. These changes
in the interfacial bond strength (with temperature)
also have a significant effect on the load transfer
characteristics between the matrix and the ribbons,
which in turn affects the failure mode of the composite
itself.‘ In order to fully understand this phenomenon,
the mechanism of load transfer from the matrix to the

ribbon needs to be understood.

The ultimate load bearing capacity of ceramic
matrix composites is usually determined by the strength
of the reinforcement (since the matrix cracks at a much
smaller strain). It is however the onset of matrix
cracking which is of significance, because it signifies
the onset of permanant damage and loss of protection
provided by the matrix against oxidation and corrosion
of the reinforcements. Hence the matrix cracking
stress is likely to be used as a design stress in
future applications. For composite systems in which
the reinforcements have the larger failure strain (and
stress), the failure mode of the matrix (and hence the

composite) is found to depend on the volume fraction of

48

the reinforcements [9]. For volume fractions of
reinforcements smaller than a "critical" volume
fraction, single fracture of the matrix is observed.
In this case failure of the matrix leads to composite
failure, since the reinforcements are unable to carry
the load when the matrix cracks. For volume fractions
of reinforcements greater than the critical volume
fraction, failure of the matrix does not lead to
composite failure since the reinforcements can still
carry the load. The failure mode of the matrix (and
hence the composite) changes to one of multiple
fracture. The "critical” volume fraction can be
obtained by a simple load balance equation and can be

written as

 

m
Vc = 1
e I + e
of of am
where, am* is the fracture strength of the matrix

of* is the fracture strength of the ribbon
I
of is the stress transfered to the ribbons

when the matrix cracks.

For large reinforcment strains (exceeding the

49

matrix failure strain), the matrix is observed to crack
into blocks of fixed lengths [10,11]. This length can
be obtained by considering the load transfer between
the reinforcements (ribbons) and the matrix. For the
geometry of the ribbon reinforcements, since the load
transfer across the ribbon matrix interface occurs by

N + —

where, N is the total number of ribbons(continuous)

r is the interfacial shear strength

a is the matrix cracking stress

Amis the cross sectional area of the matrix

w is the width of the ribbons

t is the thickness of the ribbons, and ‘

x is the length of the blocks into which the
matrix cracks(which will be equal to the
length of the ribbon initially).

Solving for x,
*

"m Am

 

2 N r (w+t)

Multiplying the numerator and denominator by wt(a constant)

50

e
"m Am "t

 

2 N r (w+t) wt

Now, Nwt is the total cross sectional area of the ribbons

(Ar) .

Dividing numerator and denominator by the cross sectional

area of the composite (Ac),

*
a m (Am/Ac) Wt

 

2 1' (w+t) (Ar/Ac)

Since the composites are continuously reinforced volume

fraction is equal to area fraction. Hence

 

V 2 ‘r (PH-t)

Since x is a function of both 0*In and (for a given
volume fraction and ribbon dimensions), the failure
mode (single matrix crack or multiple matrix crack)

will vary depending on the values of ‘fm and r.

The results of the pullout tests carried out at

fl

various temperatures are provided in Figure 4. Ribbon
pullout was observed only at temperatures exceeding
250°C. The matrix strength measured at various
temperatures is also plotted in the same figure. Some
of the interesting features that can be observed from
this figure are the following: Temperature influences
both the matrix strength and interfacial shear
strength. While the matrix strength was fairly
constant upto 400°C, the interfacial shear strength
decreased sharply at about 250°C. In an intermediate
range from 300° to 400°, the interfacial shear strength
was lower than the matrix strength. At temperatures
above 400°C, the matrix strength decreased rapidly
while the interfacial shear strength exhibited only a

moderate decrease.

These changes lead to changes in the failure
mode. Visual observations of the composite specimens
revealed that specimens tested at temperatures below
200°C exhibit a single matrix crack in composite
failure (Figure 5 a). At 300°C, the failure mode of
the matrix change to one of multiple matrix cracking
(Figure 5 b). Above 400°C, although the interfacial
strength was still decreased, the matrix strength was
decreased more rapidly due to matrix creep. Hence the

failure mode of the composite changed back to one of a

52

 

 

1 no pullout
‘3 observed
‘- A
5 :35-
f 2304 i i
V l
g 2 : viIsibl'e :
g a mu tip e
3 3.25- ‘ F—tmatrix :
a 3 I cracks :
8 a 20" : 1 matrix
.5. ‘5 n : creep
.. ‘ - I
3 5‘5 i 7‘?
5. 6‘0‘ 3 - I :‘
I. O l I
2 2 5- i :
5 t l
l l
O 1 I I I .I ' ig
0 100 200 300 400 .500

Temperature(Cl

FIGURE 4: The matrix strength and interfacial
shear strength at various temperatures.

( 0 matrix strength and A interfacial strength)

FIGURE 5:

a.
b.

c.

 

Composite failure modes at various temperatures
Single crack observed at 200°C.
Multiple matrix cracks observed at 300°C.

Single crack observed at 400°C.

54

single crack (Figure 5 c).

The effects of changes in the interfacial shear
strength on ribbon pullout can be observed in the
scanning electron micrographs provided in Figures 6 and
7. For the specimens tested at temperatures below
300°C, no pullout was visible (Figure 6 a and b) and a
strong ribbon-matrix interface was observed. At 300°C,
weakening of the interface was evidenced accompanied by
some ribbon sliding (Figure 7 a and b ). The ribbon-
matrix debonding and sliding was even more pronounced
in the specimens tested at 400°C (Figure 8). This
ribbon-matrix debonding and sliding provides for
additional energy absorption and increased fracture
toughness, at temperatures where there is no

significant loss of matrix strength.

Hence the metallic-glass reinforcements not only
improve the strength and toughness of the glass-ceramic
matrices at lower temperatures, but also provide for
high temperature toughness by means of interface
related effects such as debonding and pullout.

Although the metallic-glass ribbon/glass-ceramic matrix
combination used in the present study had relatively
lower temperature capabilities, proper choice of the

matrix and reinforcement could be used to produce

FIGURE

  
     

it
h.
“—1,
\.

            

‘v‘ j 9?} "v A?" ; V _‘ _~- ~_ ”I“

. “I - cg: ‘ ‘45:. . -. ,w ... ... ,; r_ ’.
“‘ME’f?‘ ..-. . Q~ .
fiftffigg‘? _?-.:-‘§; 1"} I‘- ’ 3. "3.; F 67 u

3:173: ‘13:)?”- v ufi'ig‘t‘t‘gj‘i 7‘4»; 1 ‘ , “fig-fl u ....— ...—4

 

Micrographs illustrating the strong

interfacial bonding for specimens tested below

300°C.

Presence of a strong void free bond between the

ribbon and the matrix.

Absence of ribbon debonding and pullout.

56

 

FIGURE 7: Micrographs illustrating ribbon-matrix
debonding and ribbon pullout for specimens
tested at 300°C.
a. Extensive ribbon debonding and pullout.
A crack is observed to initiate at the edge of
the ribbon.
b. A high magnification micrograph illustrating the

smooth matrix surface in the debonded region.

57

’EEEEEIBE (7's,
atI'ix 7.1rr

 

FIGURE 8: Micrograph illustrating extensive ribbon-matrix
debonding and ribbon pullout for the specimen

tested at 400°C.

 

58

composites with high temperature structural

capabilities.

REFERENCES:

10.

11.

Prewo K. M., J. Mater. Sci., 2 (1987) 2695.

Luh E. Y. and Evans A. G., J. Am._Ceram. Soc.,

m (1987) 466.

Rice R. N., z (1978) 69.,

Sheppard L. M., Adv. Mater. Processes incorporating
Metals progress, 139 (1966) 54.

Sheppard L. M., Adv. Mater. Processes incorporating
Metals progress, 130 (1966) 64.

Prewo K. M., Brennan J. J and Layden G. K., Am. Ceram.
Soc. Bull., 65 (1986) 305.

Vaidya R. U. and Subramanian K. N., J. Mater. Sci.,
(in press).

Vaidya R. U. and Subramanian K. N., ASM/ACCE/ESD
Conference Proceedings, Detroit, September 1989

(in press).

Hull D., "An introduction to com osite materials",
edited by Cahn R. W., Thompson M. W. and Ward I. M.,
Cambridge University press , Cambridge, (1981) pp 125.
Aveston J., Cooper G. A. and Kelly A., National
Physics Laboratory Conference Proceedings, IPC
Science and technology press, Guilford, England,
November (1971), pp 15.

Marshall D. B. Cox B. N. and Evans A. G., Acta.

Metall., 1; (1985) 2013.

59

A commercially available amorphous metal ribbon
incorporated in a glass-ceramic matrix was assessed
from the standpoint of fracture toughness. The ribbon
orientation (with respect to the long and short
transverse sides) relative to the opening crack
significantly affected the fracture toughness of such
composites. The strong dependence of the fracture
toughness on the ribbon orientation was correlated with
the differences in the bending contributions of the
ribbons in the different orientations. The effect of
the ribbon orientation on the crack growth rate in the

matrix was also investigated.

61

INTRODUCTION:

Studies carried out on metallic-glass ribbon reinforced
metal and polymer matrix composites [1-4] have
demonstrated some of the unique properties of the
ribbon reinforcements. Such composites have been shown
to exhibit high longitudional strength coupled with
good off-axis properties. Recently, the potential of
using such metallic-glass reinforcements for enhancing
the mechanical properties of brittle glass-ceramic
matrices has been demonstrated [5,6]. Ribbon
reinforcements, unlike fiber reinforcements, possess
the ribbon width as an additional geometrical
parameter. Hence the load transfer characteristics in
such ribbon reinforced composites are expected to be
influenced by the ribbon width also. Defining the
ribbon orientation in such composites is also more
involved. In addition to orienting the ribbons with
respect to the longitudinal axis of the specimen, the
ribbons can also be oriented differently with respect
to their long or short transverse faces normal to the
opening crack front (The surface bounded by the length
and width of the ribbon is referred to as the long
transverse face; one that is bounded by the thickness

and length is referred to as the short transverse

62

 

63

face). The effect of different ribbon configurations
on the fracture toughness of the composites was
investigated in this study. An understanding of this
effect would be useful in optimizing the properties of
such brittle matrix composites, in order to make them

suitable for structural applications.

 

Corning glass code 7572 was used as the starting
matrix material. METGLAS 26058-2 was used as the
reinforcement. The specimens were fabricated by
conventional wet pressing, in a steel die, using amyl
acetate as the binder. After compaction (at 3000 Psi),
the specimens were preheated to 250°C to drive off the
organic binder. The compacts were sintered at 400°C
for one hour followed by a crystallization treatment
for 20 minutes. The main crystalline phase is 2Pb0.Zn0.
Details of the physical properties and chemical
compositions of the metallic-glass and glass-ceramic
matrix are given elsewhere [6]. The residual stresses
induced by the thermal expansion coefficient mismatch
between the matrix and the reinforcement can be

neglected since a = 95 x 10-7 /°C and

matrix “ribbon =

64

76 x 10"7

/°C. Strong bonding was observed between the
ribbon and the matrix. The strong nature of the bond
was evidenced by the absence of ribbon pullout in the
pullout tests conducted [6]. A detailed study of the
nature of the bonding and interface characterization is

under progress.

Single edge notched beam specimens were used to
measure the fracture toughness. The ribbon dimensions
used were 4.0 cm x 0.5 cm x 25 microns. The total
number of ribbons (and hence the volume percent of the
ribbons) was kept the same in all the specimens (at
0.67%). The external dimensions of the composite
specimens were also kept constant (4.0 cm x 1.15 cm x
1.20 cm). The notches in the specimens were cut using
a slow speed diamond saw. The specimens were annealed
at 250°C (since the glass-ceramic has a relatively low
maximum operating temperature) after the notches were
cut, in order to heal any microcracks which might have
formed as a result of the machining operations. The
various ribbon configurations used are shown in Figure
1. The specimens were tested in three point bending in
an Instron machine with a cross head speed of 0.05
cm/min. The bending tests were performed in accordance
with A.8.T.M. STP 678 (for blunt notch specimens). 3

Five specimens were tested for each ribbon

 

'c

.L

lllllll -

 

 

 

//7///
it—bc —4

 

FIGURE 1 :

 

65

 

 

 

/77//Z?

 

 

 

 

 

 

,4666666/

 

 

 

 

 

 

 

///f///

 

 

 

1
i

 

'—%

The cross-sectional view of the different ribbon

configurations used in the Single Edge Notched

Beam specimens.

notches cut in the specimens.

The shaded area represents the

(bC is the composite width, and tc is the composite

thickness.)

66

configuration.

BEEEHI§_AED_DI§§!§§IQN=

The measured fracture toughness of the composite
specimens (along with the standard deviations) is
presented in Table 1. It is evident that the ribbon
orientation (with respect to the long and short
transverse faces) has a significant effect on the
fracture toughness of the composite specimens. The
largest increment in the fracture toughness (over that
of the matrix) was obtained for ribbons oriented with
their short transverse faces perpendicular to the
opening crack front (configuration 4 in Figure 1).

The differenCes in the fracture toughness of the
composite specimens can be explained on the basis of

three important factors.

a)- W:

The primary factor which increases the fracture
toughness in configuration 4 (as compared to the other
orientations) is believed to be the significantly

higher moment of inertia of the ribbons. Consider the

67

TABLE 1 : Fradture toughness of the composite as a

function of ribbon configuration.

 

 

Configuration RIC MPaml/2 (avg) Coeff.of Variation (8)
0 (Matrix) 0.321 7. 07
1 0.683 11.52
2 0.987 9.79
3 l. 161 5. 00
4 3. 177 16. 50

 

68

simple case of the bending of a specimen (in three
point) containing a single continuous ribbon
(lengthwise) located in the center for two different

ribbon orientations:

i). Ribbon with its long transverse face parallel to
the neutral surface and perpendicular to the opening

crack front, and

ii). Ribbon with its short transverse face parallel to
the neutral surface and perpendicular to the opening

crack front.

The moments of inertia calculated for the orientations
i). and ii). are given as I1 and Iii in equations 1 and

2 respectively.

I. =1—bt3-I-lbt3 3

1
1 6 ff TCf +1.11%;c ........ (1)

3 3

=1 3 1 1 2 _ I -
Iii -6-tfbf +_b t +thbf (tc bf) +_1.itf(tc bf) ......(2)

1200
where,
b is the width of the composite specimen
tc is the thickness of the composite specimen
is the width of the ribbon and,

69

t is the thickness of the ribbon.

f

It can be shown (for the dimensions of the ribbons
used in this study) that the bending moment required to
obtain the same maximum flexural stress in a ribbon is
about 50 times higher when the ribbons are oriented
with their short transverse side perpendicular to the
opening crack front as compared to the ribbons oriented
with their long transverse sides perpendicular to the

opening crack front.

b)- WW:

Another factor which contributes to the improved
fracture toughness in orientation 4 (as compared to
orientations 1, 2 and 3) is the differences in the
energies absorbed by the matrix. Ceramic materials can
exhibit different fracture features depending on the
velocity of the crack [7,8]. The velocity of a crack
through a brittle matrix also significanly affects the
energy absorbed by the matrix. Broberg [9] has
indicated that a substantial increase in the energy
dissipation in the plastic region (at the tip of the
crack) takes place at low crack velocities, whereas a

substantial increase in the energy dissipation in the

70

process zone occurs at higher crack velocities.

The energy absorbed by the matrix during fracture

can be written as [9]

Gcmb - R1(b-rp) + R2(rp) ................ .(3)
where,

Gcm is the critical strain energy release rate

for the matrix
b is the width of the specimen
R1 is the brittle component of Gcm

R is the ductile component of Gcm , and

2
rp is the size of the plastic zone.

For the ribbon geometry, this equation can be written

as [3]

Gem = Gm8n(As/A) + omr[1 - n(Af+As)/A]......(4)

where,
Gms is the stable (slow) crack growth component of Gcm
Gmr is the unstable (fast) crack growth component of Gcm
n is the total number of ribbons
As is the area of stable crack growth
Af is the cross sectional area of a ribbon, and

A is the cross sectional area of the composite

N

specimen.

Since the energy required for propagating an
unstable crack is greater than the energy required for
propagating a stable crack in a ceramic material [7,8],

one can assume

%r=x%s ”HHHHHHHHH ........ a)

where, K is a positive constant, greater than one.

Using equation (5), equation (4) can be
simplified to a form

%m=%JC-D%]Hnunnununnw)

1’.—

where C and D are positive constants. As per this
equation, Gcm is inversely proportional to the area of

stable crack growth.

Previous studies carried out on brittle ceramic
materials [7,8] have indicated that it is possible to
distinguish between the areas of stable and unstable
crack growth using microscopic techniques. In the case
of stable crack growth (in the brittle matrix), the

crack has enough time to choose its path, and

72

propagates along a path of least resistance (usually
grain boundaries). Thus the fracture for a specimen in
which the crack has grown slowly is predominantly
intergranular. In case of unstable (rapid) crack
growth, the crack grows in a linear self-similar
manner, and the fracture observed in this case is

predominantly transgranular.

Fels et a1. [3] have considered the effect of
ribbon orientation on the velocity of the crack through
the matrix. A similar analysis will predict the area of
stable crack growth in configuration 4 to be
significantly smaller than that in the other three
configurations. This is because of the smaller crack
front intercepted by the ribbon. The smaller dimension
of the ribbon intercepting the rapidly propagating
crack front decelerates only a small portion of it. As
a result the area of stable crack growth in the matrix,
which is located in the vicinity of the ribbon is
reduced dramatically. Consequently only a very small
portion of the crack front needs to accelerate as it
leaves the ribbon to catch up with the rapidly
propagating front. In this manner, the fracture
behaviour of the matrix, influenced by the presence of
the ribbons, also contributes to the increased fracture

toughness.

73

In order to observe the regions of stable and
unstable
crack growth in the system under investigation, the
composite specimens were lightly etched with a 3%
hydroflouric acid solution. Scanning electron
micrographs (Figures 2 a and b) obtained for specimens
containing ribbons in the two different orientations,
reveal some distinctive features. In both cases
(ribbon oriented with its short or long transverse
surfaces perpendicular to the crack front), the ribbon
arrested the rapidly propagating matrix crack. The
crack could not bypass the ribbon, because the crack
has to intersect the ribbon since it will propagate
along a plane perpendicular to the length of the
ribbon. Once the crack propagates through the ribbon,
it propagates catastrophically (because of the large
amount of energy available for its growth) and tries to
catch up with the the rest of the crack front that has
propagated in regions in which it has not been
intercepted by the ribbon. The strong ribbon-matrix
interface prevents crack deflection at the interface.
The region adjacent to the ribbon indicated the
presence of an ”intermediate zone" (marked ‘X'in
Figures 2 a and b) , wherein the crack accelerated to
catch up with the propagating crack front in the other

regions of the matrix. The crack usually initiates at

74

FIGURE 2:a. Scanning electron micrograph illustrating the
fracture features observed for a specimen
containing ribbons oriented with their long
transverse faces perpendicular to the crack.
front.

b. Scanning electron micrograph illustrating the
stable and unstable crack growth regions in a
specimen containing ribbons oriented with
their short transverse faces perpendicular

to the crack front.

Arrow indicates the direction of crack propagation
in the composite. ‘X' and ‘Y' correspond to
~regions of unstable and stable crack growth
respectively. ‘X' corresponds to the intermediate zone

referred to in the text.

75

 

76

a different level on the far side of the ribbon
(probably due to flaws existing at the interface, or
due to bending of the ribbon), and accelerates to catch
up with the propagating crack front. [In doing so, the
crack has to climb, thereby traversing a longer and
tedious path. However, this region (intermediate zone)
appears to be smoother as compared to the surrounding
region (marked ‘Y' in Figures 2 a and b). This
indicates that the crack propagates more rapidly in
this region. The size of the intermediate zone depended
on the ribbon orientation. This zone was relatively
larger in the specimens containing ribbons oriented
with their short transverse faces perpendicular to the
crack front.

This observation is not in agreement with the
model of crack growth in such a region proposed by Fels
et a1. [3] for polymer matrix composites. This is
probably as a result of the different magnitudes of
interfacial bonding between the component phases. In
the case of the system under investigation, the strong
bond between the matrix and ribbons prevented any
matrix sliding. Hence the crack cannot propagate
through the matrix without fracturing the ribbons. In
case of the metallic-glass/polymer matrix composite

studied by Fels et al. [3], the interface was

relatively-weak; Hence the crack could propagate

77

through the matrix without ribbon failure, although at
a much slower rate. Similar features were observed in
studies on a nichrome ribbon/slide glass system [10],

where the interfacial bond strength was relatively low.

c) -_Eibmn_enasins_effeets=

The differences in the fracture toughnesses
measured for configurations 1,2 and 3 could be
attributed to the different spacing of the ribbons
(since the moments of inertia for all three
orientations are the same). The spacing between
ribbons in configuration 1 is much smaller than the
spacing between ribbons in configurations 2 and 3.
Hence, for configuration 1 it is possible that the
intermediate zone overlaps with ribbons, thereby
absorbing a smaller amount of energy. The differences
in the spacing of the ribbons in configurations 2 and 3
is small. In fact configuration 3 has a slightly
smaller ribbon spacing as compared to configuration 2,

but exhibits a higher K This is probably due to the

Ic‘
more complex arrangement of the ribbons leading to a
larger intermediate zone. In addition, the propagating
crack front will be intersected by the ribbons more

effectively in this ribbon configuration. Although the

78

presence of stable and unstable crack growth was
detected in the matrix, further studies to quantify the

process are essential.

.QQNQDQSIQHE3

The fracture toughness of continuous
metallic-glass reinforced glass-ceramic matrix
composites is strongly dependant on the ribbon
orientation (long transverse or short transverse
faces) with respect to the opening crack front. This
was attributed to the differences in the bending
moments associated with the different ribbon
orientations. The ribbon orientation was affected the
crack velocity in the matrix and subsequently the

fracture toughness of the composites.

AEIEQELEDEEHEEI§=

The authors would like to acknowledge the
Composite Materials and Structures Center at Michigan
State University for supporting and funding this

project. The authors would also like to thank Dr. N.

79

J. Altiero of Michigan State University for his helpful
suggestions during the preparation of this manuscript.
The authors would also like to extend their sincere
thanks to Dr. Kenneth Chyung of Corning Glass Inc. and
Dr. Edward Norin of Metglas Products for their valuable
help and suggestions during the early stages of this

project.

REFERENCES:

Cytron S. J., "A metallic glass metal matrix composite"
J. Mater. Sci. Letters., _l_(1982) 211-13.

Strife J. R. and Prewo K. M., "Mechanical behaviour

of an amorphous ribbon reinforced resin matrix composite",
J. Mater. Sci., _11 (1982) 359-68.

Fels A., Friedrich K. and Hornbogen E.,"Reinforcement

of a brittle epoxy resin by metallic glass ribbons",

J. Mater. Sci. Letters., _; (1984) 569-74.

Friedrich K., Fels A., and Hornbogen E.,"Fatigue and
fracture of metallic glass ribbon/epoxy matrix
composites", Comp. Sci. and Tech., _;; (1985) 79-96.
Vaidya R. U. and Subramanian K. N., "Metallic-glass
reinforcement of glass-ceramics" ASMzACCEzESD Conference
Proceedings, Detroit, Published by American Society of
Metals, Metals Park, Ohio (1989) pp 227-231.

Vaidya R. U. and Subramanian K. N., "Metallic-glass
ribbon reinforced glass-ceramic matrix composites",

J. Mater. Sci. (in press).

Jessen T. L. and David Lewis III, "Effect of crack
velocity on crack resistance in brittle-matrix
composites", J. Am. Ceram. Soc., 1;[5] (1989) 818-21.
Jessen T. L., Mecholsky J. J. and Moore R. H., "Fast

and slow fracture in glass composites reinforced with

80

10.

M

Fe-Ni-Co alloys", Am. Ceram. Soc. Bull.,

65(2) (1986) 377-81.

Broberg K. B., "On the effects of plastic flow at fast
crack growth" from W. ASTM
STP 627, edited by Hahn G. T. and Kanninen M. F.,

American Society for Testing and Materials, 1977 pp 243-56.
Vaidya R. U., Lee T. K. and Subramanian K. N., "Metallic
ribbon reinforcement of ceramics" ASC Fifth Technical

Conference Proceedings. (in press).

82

AB§IBA£I=

The load transfer mechanism inta metallic-glass
ribbon reinforced glass-ceramic matrix composite was
investigated. The critical ribbon length required for
effective load transfer between the matrix and the
ribbons, was found to be a function of the ribbon
width. This dependance was attributed to the size of

the long-transverse side available for load transfer.

83

Unidirectional fiber reinforcement has been
found to be a very effective method for improving the
strength of brittle glass and glass-ceramic matrices
[1-3]. The major drawback of such unidirectionally
reinforced composites is their relatively poor off-axis
properties. One of the ways in which this problem has
been overcome is in the use of multiple plys,
containing fibers oriented at different angles (0°,

°, 90° etc) with respect to the longitudinal axis

45
[4]. Although such composites are isotropic on a
macroscopic scale, the properties within each ply are
not the same in different directions. This anisotropy

within each ply can have serious consequences on the

failure mode of such composites.

Ribbon reinforced composite plys have been found
to exhibit transverse strengths approaching about 50%
of their longitudinal strength. These are achieved
because of the unique geometry of the ribbon
reinforcements. Metallic-glasses appear to be the best

candidates as ribbon reinforcements for the following

84

85

reasons. They exhibit very high strength and fracture
toughness (far exceeding that of the parent metal from
which they are derived). Their ribbon geometries are
also a natural consequence of the extremely high
cooling rates required in their manufacture. Metallic-
glass reinforced polymer and metal matrix composites
have been found to exhibit very high strengths [5-7].
Recently, metallic-glass ribbons were investigated as
reinforcements for brittle glass-ceramic matrices.

Even small volume fractions of such reinforcements
improved the mechanical properties of the glass-ceramic

matrices quite significantly [8]. g

The geometry of the ribbons is significantly
different from that of conventional fibers, and as a
result the load transfer characteristics are expected
to differ from those related to fiber reinforced
systems. An understanding of this feature is important
for structural applications of ribbon reinforced
composites. The present study addresses some of the
aspects of load transfer in a metallic-glass ribbon,

reinforced glass-ceramic matrix composite.

133981;

In the case of metallic-glass reinforced glass-

ceramics, where both the failure stress and failure

strain of the reinforcement exceed that of the matrix,

the composite strength can be expressed by

*_ t 1 f
60 - om Vi + of Vf or Vf < Vc ........(1)
d * - *v f v > v 2
an ac - of f or f c eoooeeoeoe()

where, “c is the
* 0
“m is the
of* is the
I
of is the
matrix

composite fracture strength
matrix fracture strength
ribbon fracture strength
stress in the ribbons when the

is about to fracture

Vf is the volume fraction of the ribbons

Vt is the volume fraction of the matrix , and

Vc is the "critical volume fraction" of the ribbons.

The ”critical volume fraction" is the minimum

volume fraction of reinforcements required for

preventing the catastrophic failure of the composite

when the matrix fractures.

combining equations

It can be obtained by

1 and 2.

86

Hence,

 

"m
Vc = * oeeeoeeee(3)
* 1 + *
"f .°f “m
For the system used in this investigation, °m* =
I
15 MPa * = 700 MPa and = 685 MPa. The critical

”f ”13
volume fraction for the given system was determined

0
using these values and was found to be 0.5%.

When the composite is loaded, the failure strain
of the matrix is reached first and it cracks. The load
. is transferred to the ribbons by shear across the

interface. Hence,
*
‘mAm = 27(wx+xt) ..... (4)

where, 1' is the interfacial shear strength
Am is the total area of the matrix.
x is the ribbon length
w is the ribbon width, and
t is the ribbon thickness.

This equation can be rewritten in terms of the

volume fractions of the ribbons and the matrix as

2 1(wx + xt)Vf

 

a V "-" ............. (5)

For the system under investigation, the Young's
modulus of the ribbons is greater than that of the
s 85 GPa and E

matrix (E = 33 GPa).

ribbon matrix

In the initial stages of loading (for a volume
fraction of reinforcements greater than the critical
volume fraction), the load is carried by the matrix and
the ribbons. When the failure strain of the matrix is.
reached, the matrix fails and the load is transferred
to the ribbons. The ribbons continue to carry the load
till failure.

The maximum strength of the composite is reached
if the ribbon length is greater than a critical length
*
BC when of = of . Hence (for x = 9c) the maximum

composite strength is

2 1 Vf(w +t)9c

 

v = .......(6)

 

In order to determine the critical load
transfer length, ribbons of varying lengths and three
different widths (0.2, 0.5 and 0.8 cm) were
incorporated in the glass-ceramic matrix. The specimen
length ( 4.0 cm), the specimen width ( 1.0 cm) and the
volume fraction of the ribbons (0.67%-greater than the
critical volume fraction) were maintained constant for

all the specimens.

Metglas 26058-2 was used as the reinforcement.
Corning glass code 7572 was used as the starting
matrix material. The specimens were fabricated by
conventional wet pressing, using amyl acetate as the
binder in a steel die with a pressure of 3000 Psi.
After compaction, the specimens were preheated to 250°C
to drive off the organic binder and then sintered at
400°C for one hour. The crystallization of the matrix
was achieved by holding the composite at 450°C for 20
minutes. Details of the physical and chemical
properties of the metallic-glass reinforcement and
glass matrix used are given elsewhere [8]. Residual
stresses induced by the thermal expansion coefficient
mismatch between the matrix and reinforcement ((1

matrix

- 95 x 10-7/ °C and (1 s 76 x 10-7/ °C) can be

ribbon

90

considered to be negligible.

Due to the difficulties encountered in gripping
of tensile specimens, the strength of ceramics is
normally evaluated using bending tests. The maximum
outer fiber stress (usually expressed as the Modulus of
Rupture, MOR) is taken as a measure of strength. In the
present study, tests were carried out in three point
bending in accordance with ASTM standard No. C-203/85.
All the specimens were carefully polished (using 0.03
micron alumina powder) before testing in order to
minimize the effect of surface flaws. A total of five
specimens were tested to obtain the average value for
each data point. In all the specimens, the failure
occurred at the midpoint, which experiences the maximum

bending moment.

BEEQHT§_AND_DI§§Q§§IQN=

The variation in the strength of the composite
specimens as a function of ribbon lengths, for three
different ribbon widths are shown in Figure 1. Similar
trends have also been observed in other fiber

reinforced systems [9].

91

 

 

 

f‘
G
O.
2 40
v T
I: 35-
:5 M»
5-
0 2
h
3 20..
a
CL
3 15-
0:
10-
u—
o A ribbon width 0.2 cm
U3 5‘ I ribbon width 0.5 cm
a O ribbon width 0.8 cm
- l l I
:5 00 i 2 3 4
13
§

ribbon length L (cm)

FIGURE 1: Variation in the composite fracture strength

as a function of ribbon length.

92

Increasing the ribbon length enhances the load
transfer upto a certain "critical" ribbon length.
Increasing the length of the ribbon beyond this
”critical” length does not improve the composite
strength. For a critical ribbon length 2c,
corresponding to a ribbon width of 0.8 cm (the maximum
based on experimental limitations), the composite

strength can be approximated by the equation

2 1Vf(w + t)

 

.C 2,, epr-(szc-xi/xm + (am*vm>....(7)

where the exponential factor accounts for the load
carrying capacity of the composite for ribbon lengths
in the range of 0 < K < 2c. This exponential factor
was obtained by emperical curve fitting. This equation
satisfies the boundary conditions. When x = 0, the

ribbon contribution is zero and the composite strength

is equal to the matrix strength. For x QC (critical
length), the composite strength is a maximum as given

by equation 6.

Studies carried out on other metallic-glass

ribbon reinforced systems (polymer matrices) [7], have

93

assumed the load transfer from the matrix to the
reinforcing ribbons to be linear. According to this

study, the composite strength can be given by

“c = .f*vf(1 - Qt/Q) + ( ,m*vm)...........(8)

where Qis the ribbon length and Qt is the critical
transfer length. However, the experimental findings
for the system under investigation did not agree with
this equation. In addition, this equation can only be
used to predict composite strengths for ribbon lengths
in the range 0 < 2 < 2t. It does not satisfy either of

*

the limiting conditions (at Q=0, ac = °m*vm and at?

* t
8 Qt] 0c ‘ of vf)
Since the thickness of the ribbon is very small
as compared to its width, (w + t) can be approximated

to ‘w'. Hence equation 7 can be simplified to

21'vf Qc

= exp[-(Qc-x)/x] + ( am*Vm).....(9)

 

Taking the logarithms of both sides of equation

9 and simplifying

94

*
ln( °c* - ”m Vm) = K - QC/x ............ (10)
where K is a positive constant.

A plot of ln( ,C*

in Figure 2. The slope of this line is the critical

* -1 o
- om Vi) versus X is shown
length 9c, which was found to be 2.581 cm.

Although the equations described were developed
for a ribbon width of 0.8 cm, the same equation can be
used for composites containing ribbons of other widths,
provided a geometrical correction factor is applied.
This correction factor can be referred to as the ”shift

factor".

The concept of ”shift factor" can be understood
by considering the experimentally obtained data of the
composite strength as a function of ribbon length
(Figure 1). This strength data can be replotted
against the ribbon length to width ratio (Figure 3).
The strength curve appears to shift depending on the
ribbon width when plotted against the ribbon length to
width ratio. The curve corresponding to a ribbon width
of 0.8 cm was chosen as the master curve. All the data
corresponding to the other ribbon widths can be shifted

onto this master curve by using the "shift factor".

 

95

 

 

 

 

 

1.1

FIGURE 2: Variation in the composite fracture strength
as a function of reciprocal ribbon length.
Strength is expressed in MPa, and reciprocal

length is expressed in (cm)-1.

96

 

 

 

 

critical ribbon length Jc(cm)

 

2 . I ‘ l ' I I
02 ' 04 06 08 t0

ribbon width w(crn)

FIGURE 3: Variation in the critical ribbon length as a

function of ribbon width.

 

86286188

composite bending strength (MPa)
01

0

97

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

012 8 4 5 8 7 8 91011121814151817181920
ribbon '9091h/widthU/w)

_a—w 0.2 cm —+—w 0.5 cm 44w 0.8 cm

FIGURE 4: Variation in the composite bending strength

with the ribbon length to width ratio.

The shift factor "a" can be defined as

length of ribbon required to obtain a given
strength in a composite containing ribbons of

width 0.8 cm

 

length of ribbon required to obtain the same
strength in a composite containing ribbons

of width other than 0.8 cm

According to this definition

a = 1 for a ribbon width of 0.8 cm, and

a > 1 for a ribbon width less than 0.8 cm.

Values of the shift factor can be obtained from
the plot of the strength versus the ribbon length to
width ratio (Figure 3). Variation in the reciprocal

shift factor with ribbon width is shown in Figure 4.

The shift factors can now be used to shift all
the data corresponding to various ribbon widths onto a
single master curve (Figure 5), facilitating the use of
the earlier proposed equations for the composite

strength.

99

A

U
\1_
,_
v4
L.

C) '1
.+3
04
O
H—e
£4
.3: -+
U)

.L 1L.

 

 

recuproca
CD

I T I r If I I r T

ribbon width w (cm)

FIGURE 5: Variation in the reciprocal shift factor

with ribbon width.

 

100

 

 

 

O 1 2 3 4-
normalized ribbon length (l/a)

composite strength (MPa)

FIGURE¢5: Variation in the composite bending strength

with normalized ribbon length.

101

The ribbon width was found to have a significant

effect on the load transfer between the ribbon and the
matrix. A plot of the variation in the critical length
as a function of ribbon width is shown in Figure 6.
This effect is probably a consequence of the non-
uniform bending moment along the specimen and hence the
non-uniform stress in the cross-section experienced by
the composite specimen in the bend test. In the three
point bending configuration the bending moment is a
maximum in the center of the specimen and drops of to
zero at the specimen ends. The specimen failure is
usually initiated in a region close to where the
maximum bending moment (and hence the maximum outer
fiber tensile stress) is encountered. The effective
ribbon area encountered by a crack decreases as the
ribbon width decreases. This ribbon area is critical
because stress transfer occurs from the matrix to the
ribbon across this area. This explains the larger
critical lengths observed for the specimens with the

smaller width reinforcements.

Another factor which localizes the stress to a
very small region in the matrix is the presence of an
extremely strong bond between the ribbon and the
matrix. The nature of the bonding between the matrix

and the reinforcements can be seen in Figures 7 a and

102

FIGURE 7: Micrographs illustrating the nature of the
bonding between the glass-ceramic matrix

and the metallic-glass ribbon reinforcement.

a. Absence of ribbon pullout in a pullout test.

b. Presence of a strong void-free bond between

the ribbon and the reinforcement.-

 

103

 

104

b. According to the ACK theory [10], the interfacial
strength between the reinforcement and matrix controls
the load transfer across the interface and also affects
the load transfer length. The load transfer length can

be obtained by rearranging equation 5 and solving for

‘x'. Hence,

.............. (11)

 

21Vf(w+t)

According to this equation a strong interface
results in the load being transferred across a very

narrow region in the center of the specimen.

Conclusively it can be said that the load
transfer in ribbon reinforced composites cannot be
characterized by a simple aspect ratio of length to
‘thickness as considered with other fiber reinforced
systems. Specifying the width in such composites is
eessential, since it affects the load transfer between

the matrix and the ribbons quite significantly.

CONC S ONS:

The present study indicates that the load
tranfer characteristics in ribbon reinforced composites
differ from those in fiber reinforced systems. It is
not possible to characterize the load transfer in such
composites by a simple aspect ratio (i/d) as in fiber
reinforced systems. The width of the ribbon plays a
significant role in the load transfer process. The
interfacial bond strength is also found to
significantly affect this dependency. Further studies

are needed to completely characterize this behaviour.

105

The authors would like to acknowledge the
Composite Materials and Structures Center at Michigan
State University for funding and supporting this

project.

NCES:

1. Fitzer E. and Gadow R., Am. Ceram. Soc. Bull.,
g; (1986) 326.

2. Herron M. and Risbud 8., Am. Ceram. Soc. Bull.,
65 (1986) 342.

3. Prewo K. M., Am. Ceram. Soc. Bull., g; (1986) 395.

4. K0 F., Am. Ceram. Soc. Bull., 61 (1989) 401.

5. Cytron S. J., J. Mater. Sci. Letters, 1 (1982) 211.

6. Strife J. R. and Prewo K. M., J. Mater. Sci., 11
(1982) 359.

7. Strife J. R. and Prewo K. M., AFWAL Report
AFWAL-TR-80-4060, 1980.

8. Vaidya R. U. and Subramanian K. N., J. Mater. Sci.,
(in press).

9. Hancock P. and Cuthbertson R. C., J. Mater. Sci., 5
(1970) 762.

:10. Aveston J., Cooper G. A. and Kelly A., National

Physics Laboratory Conference Proceedings, IPC

Science and Technology Press, Guildford, England,

November 1971, pp 15.

107

CHAPTERS

108

AB§IEA£I=

Discontinuous metallic-glass ribbons of varying
lengths and widths were used to reinforce a brittle
glass-ceramic matrix. The fracture strength and
toughness of such composites as a function of ribbon
volume fraction and geometry were measured in three
point bending. The mechanical properties were found to
be relatively isotropic in the plane of compaction
(without significant loss of strengthening achieved
with unidirectional reinforcement). The higher
composite strength exhibited in a direction
perpendicular to the plane of compaction was attributed
to the higher percentage of ribbons oriented with their
short transverse faces perpendicular to the opening

crack front.

109

IBIEQDQQIIQN=

The potential of using metallic-glass ribbons as
reinforcements for brittle glass-ceramic matrices has
been demonstrated [1-5]. Such composites were produced
by relatively simple, and cost effective pressing and
sintering techniques. The studies carried out so far
have indicated that significant improvements in the
mechanical properties of the glass-ceramic matrices can
be achieved by the use of relatively small volume

fractions of metallic-glass ribbon reinforcements.

Some of the advantages of metallic-glasses which
make them attractive as reinforcements is their high
strength, toughness and corrosion resistance. The high
cooling rates'required in their manufacture imparts the
ribbon geometry. This ribbon geometry not only
(provides for a large surface area to bond with the
matrix, but has also been found to be useful in
:1mproving the off-axis properties, when used as
Jaeinforcements for brittle polymer matrix composites

[6,7].

The studies carried out so far [1-5] have

foccused on the reinforcement of brittle glass-ceramic

H0

H1

matrices with continuous metallic-glass ribbon
reinforcements. However, these continuously reinforced
composites are unsuitable for structural applications,
because of the high degree of directional strengthening
they provide. Most structural applications demand a
good degree of strength isotropy. Fabrication of these
continuously reinforced composites is relatively
difficult (since they involve ribbon lay up etc). The
extremely small thickness of the ribbons also makes it
difficult to incorporate large volume fractions of
reinforcements into the matrix. In order to overcome
these problems, discontinuously reinforced metallic-
glass ribbon/glass-ceramic matrix composites were
fabricated. The mechanical properties of these
composites were measured and compared with those of
continuously reinforced composites. The load transfer
behaviour in such composites was also studied, and

compared to the theoretical model proposed earlier [8].

 

Corning glass 7572 and METGLAS 26058-2 were
used as the matrix and reinforcement respectively.
Details of the physical properties and chemical

compositions of the matrix and reinforcement are

112

provided elsewhere [1].

Three different ribbon dimensions were used in
the study viz. 0.5 cm x 0.5 cm (length x width), 1.0 cm
x 0.5 cm and 1.0 cm x 0.25 cm. The ribbon thickness
was kept constant (25 micrometers). The ribbons were
precoated with the glass powder prior to composite
fabrication. The precoated reinforcements were mixed
with the glass powder (containing 8-10% by weight of
amyl acetate binder). The ribbons were distributed as
randomly as possible so as to prevent the ribbons from
orienting preferentially with respect to one particular
face (Figure 1). Pressing was carried out at 3000 psi
in a steel die. After compaction, the specimens were
heated to 250°C for 4-5 hours to drive off the organic
binder. Sintering was carried out at 400°C for 3 hours
and was followed by a crystallization step at 450°C for
30 minutes in order to convert the matrix into a glass-

ceramic.

The specimens were cut from the fabricated
rectangular bars, and were tested in three point
bending in an Instron machine, using a cross-head speed
of 0.05 cm/min. The three point bend tests were
carried out in accordance with ASTM Standard C-203/85.

Fracture toughness measurements were carried out using

113

longitudional

   
  
 

direction

long transverse
taco

short transverse lace

edge taco

FIGURE 1: Illustration of a ribbon reinforcement
indicating the terminology used to represent

the various surfaces and directions.

H4

Single Edge Notched Beam specimens (in three point
bending). The tests were carried out in accordance
with ASTM STP 678. A total of three to five specimens
were used to obtain the average value for each data
point. In all the specimens failure occurred at the

midpoint.

BE§ELE§_A!D_DI§£!§§IQE=

The results of the experimentally measured
bending strength as a function of increasing volume
fraction of reinforcements (for different ribbon
dimensions) are given in Table 1. From the data it is
evident that increasing the volume fraction of
reinforcements increases the strength. The
experimentally measured strength of the discontinuous
composites was about 80% of the strength of the
continuously reinforced composite having the same
volume fraction of reinforcements. The strength of the
discontinuously reinforced composites can be compared
to the theoretical composite strength (continuously

reinforced) in Figure 2.

The strength of the composite was measured in

three different directions; longitudional denoted by

H5

TABLE 1: Experimentally measured composite strength as
a function of ribbon volume fraction, for three

different ribbon dimensions.

 

Volume fraction Ribbon dimensions (l/w) Average Composite 90011-01

 

of ribbons (%) length 1 width w fracture strength Vhri‘t100 (*’

(cm) (an (“MD

5.8 0.5 0.5 1 127.75 6.8

5.8 1.0 0.5 2 141.99 8.4

5.8 4.0' 0.5 8 162.4

5.8 1.0 0.25 4 132.12 10.5

3.5 0.5 0.5 1 91.5 3 7

3.5 4.0' 0.5. 8 98.0

2.5 0.5 0.5 1 58.37 2.7

2.5 1.0 0.5 2 73.41 6.9

2.5 4.0' 0.5 8 70.0

2.0 0.5 0.5 1 47.18 5.5

2.0 4.0* 0.5 8 56.0

2.0 1.0 0.25 4 55.0 11.8

0 indicates ribbon length for a continuously reinforced

composite, based on maximum die size.

 

Modulus of Rupture MOR (MP0)

1801
170-
160-
150-
140—
130-
120-
110-
100—
90-
80—
70-
60—
504
40—
30-
20-
10..

116

 

 

 

l l l l 1

1 2 3 4 ' 5
Volume fraction of ribbons Vf (7.)

ma

H7

FIGURE 2: Plot of the composite bending strength as a
function of increasing volume fraction of
metallic-glass reinforcements.

The solid line drawn through (.) indicates the
theoretically predicted composite strength
(using 0

c
represent three different ribbon dimensions as

* =°f*vf)' The other legends

(II) length 0.5 cm, width 0.5 cm
(A) length 1.0 cm, width 0.5 cm

(O) length 1.0 cm, width 0.25 cm.

Ribbon thickness was 25 micrometres.

H8

‘L', transverse denoted by ‘T' and in a direction
perpendicular to the plane of compaction denoted by
‘CT'. These three orientations are shown in Figure 3.
All the specimens were cut from the same block of .
composite material in order to avoid any discrepancies
induced due to processing. The strength values obtained
for the three different orientations are given in Table
2. The strength of the composite in the transverse
direction (T) is about 95% of that in the longitudional
direction (L), which in turn is about 80% of the
strength of a continuously reinforced composite in the
longitudional direction. From the strength values
given in Table 2 it is evident that the composite
exhibits strength isotropy in the plane of compaction.
These results are illustrative of the advantages
provided by ribbon reinforcements over fibers.
Continuous fiber reinforced composites are extremely
weak in the transverse direction (5-25% of the
longitudional strength). This drawback is overcome by
using multiple plies (containing ribbons oriented in
different directions), but at the expense of a
reduction in the strength in the longitudional
direction. On the other hand, continuous ribbon
reinforced composites have been shown to exhibit
transverse strengths as high as 50 % of their

longitudional strengths [7]. By using discontinuous

H9

compaction
direction

 

 

 

 

 

 

l7: fil/‘er

. i
-/- Compact dimensions: 1 = 4.0 cm
‘ , W = 2.5 cm
’7 " t = 0.8 cm

FIGURE 3: Figure illustrating the various orientations in

which the composite specimen was tested.

120

TABLE 2: Experimentally measured composite strength in
three different directions. The three

directions used are identified in Figure 2.

Volume fraction of ribbons 5.8%
Dimensions of ribbons length 0.5 cm
width 0.5 cm

thickness 25 micrometers

Theoretical composite strength 162.4 MPa

 

 

Direction Average Composite 90011-01 VirittIOH
fracture strength (MPa) (%) I
L 132.3 5.8
T 125.8 10.1

CT 146.2 3.8

 

1N

ribbon reinforcements of appropriate dimensions, an
even higher degree of strength isotropy can be

obtained.

In addition to the ribbon geometry, the strength
of the bond between the ribbon and the matrix is also
significant. Studies carried out on metallic-glass
ribbon/polymer matrix composites (which exhibit
relatively weak interfacial bonding) [6,7], and
Nichrome ribbon/slide glass matrix composites [5] have
indicated a lower degree of strength isotropy as
compared to the system investigated. Pullout tests
carried out on the metallic-glass/glass-ceramic system
have indicated the presence of a very strong bond
between the matrix and the ribbons. The nature of this
band can be observed in Figure 4. This strong nature
of the bond is primarily responsible for the high level
of strength utilization of the metallic-glass ribbons,

and also the isotropy in the system.

Thus the high transverse strength and in-plane
isotropy in such metallic-glass ribbon/glass-ceramic
matrix composites can be attributed to three important

factors:

122

. . 7

.5 I‘lb on

ig’? [am 13%
”I

6 .

fi.

.13" I. “_ .

.. (E.,; 1%.
x

‘0—
I

.49

I
o
I-
l

 

FIGURE 4: Scanning micrograph illustrating the strong

bond between the matrix and the ribbon.

123

1. High strength of the metallic-glass ribbons
2. Unique ribbon geometry, and
3. Good bonding between the metallic-glass ribbons and

the matrix.

Specimens with the ‘CT' configuration exhibited
the maximum strength. Although the ribbons were
distributed as randomly as possible, a larger fraction
of the ribbons reoriented themselves with their long
transverse faces (Figure 1) perpendicular to the
direction of compaction during the compaction and/or
sintering stage. Scanning microscopy carried out on
the fractured surfaces confirm this effect (Figure 5).
The higher strength obtained for this orientation can
be primarily attributed to the higher moment of inertia
of the ribbons in that orientation. Details of the
orientation effects in such metallic-glass
ribbon/glass-ceramic matrix systems have been discussed

elsewhere [4,5].

The results of the fracture toughness tests
carried out are presented in Table 3. The fracture
toughness of the composite specimens was significantly
isotropic (though slightly lower in the transverse

direction), and could possibly be attributed to the

124

 

FIGURE 5: Micrographs of the discontinuously reinforced
composites illustrating a large fraction of the
reinforcements oriented with their long
transverse faces perpendicular to the

direction of compaction

H5

TABLE 3: Experimentally measured composite fracture

toughness in two different directions. These

directions are identified in Figure 2.

Volume fraction of ribbons 5.8%
Dimensions of the ribbons length 0.5 cm
width 0.5 cm

thickness 25 micrometers

 

Direction Average composite fracture Coeff.of Viriation

toughness KIc (MPa ml/Z) (%)

 

 

H6

non-uniform ribbon distribution.

In order to characterize such discontinuous

metallic-glass ribbon reinforced composites, it is

necessary to examine the load transfer mechanism in

such systems. Previous studies [8] carried out on such

composites have indicated that both the ribbon length

and width affect the load transfer in such composites.

The two important equations developed as a result of

that analysis are

2 r Vf(w + t)Qc

 

c
wt
and
* = *
“cm °f Vf
* .
where, “c is the
*
“cm is the
of* is the
4
am is the

Vf is the

eXPE-(Qc-X)J + ( am*vm)

2 er(w + t) QC

 

composite fracture strength
maximum composite strength
ribbon fracture strength
matrix fracture strength

volume fraction of the ribbons

EN

V is the volume fraction of the matrix

x is the ribbon length

QC is the critical ribbon length (for this
geometry QC = 2/w) ‘

w is the ribbon width, and
is the ribbon thickness.

These equations were developed assuming perfect
bonding between the matrix and the reinforcements, and
assuming load transfer from the matrix to the ribbons
occurred by shear. The exponential term in Equation 1
was obtained by experimental data fitting and
satisfying the boundary conditions: for x = 0 (no

ribbon), * = *Vm' and for x = 9c, a * = a *.

aC om
According to Equation 1, increasing the ribbon

length (till 9c) causes an increase in the strength of
the composite. The maximum composite strength is
obtained for a ribbon length of QC (Equation 2). The
critical ribbon length depends on the width of the
ribbon. Increasing the ribbon width decreases the
critical ribbon length required for maximum load

transfer.

A comparison between the theoretically predicted

strength and experimental results can be seen in Figure

128

6. A, B, C, D, E and F represent curves for different
volume fractions and critical lengths of
reinforcements. These curves were plotted by using
Equation 1. From this equation it can be seen that the
maximum load transfer for a given ribbon width is
obtained for the same critical ribbon length. However,
the maximum load carried depends on the volume fraction
of the reinforcements (and increases with increasing
volume fraction of reinforcements). A, C, D and E
represent curves plotted by using the critical length
corresponding to a ribbon width of 0.5 cm, and
different volume fractions of ribbon reinforcements. B
and F represent curves corresponding to a ribbon width
of 0.25 cm (but with different volume fractions of
ribbon reinforcemnets). It can be seen that the ribbon
width affects the critical ribbon length (compare
curves A with B and curves E with F) but not the
maximum load carried. The experimental results obtained

are in agreement with the theoretically proposed model

[8].

The studies have indicated the feasibility of
developing discontinuous metallic-glass ribbon
reinforced glass-ceramic matrix composites with
structural capabilities. Although some information on

the load transfer in such composites has been obtained,

129

FIGURE 6: Plot of the composite strength as a function of

the ratio of ribbon length to width.

Solid curves A,B,C,D,E and F are obtained by

using equations

1 and 2.

(Vf is the volume percent of the ribbon

reinforcements and w is the ribbon width.)

A : Vf = 5.8% and w
B : Vf = 5.8% and w
C : Vf = 3.5% and w
D : Vf = 2.5% and w
E : Vf = 2.0% and w
F : Vf = 2.0% and w

0.5 cm

= 0.25 cm

= 0.5 cm

= 0.5 cm

= 0.5 cm

= 0.25 cm

The experimentally measured strength values are

indicated by various

(0) Vf = 5.8% and w
(7) Vf = 5.8% and w
(A) Vf = 3.5% and w
(II) Vf = 2.5% and w
(O) Vf = 2.0% and w
(as) Vf = 2.0% and w

legends as;

= 0.5 cm
= 0.25 cm
= 0.5 cm
= 0.5 cm
= 0.5 cm

= 0.25 cm

Modulus of Rupture MOR (MPG)

1805
170-
160-
150.-
140~
130-
120-
1104
1004
90—
804
70-
60—
50—
40-
30-
20-
10-

 

B0

 

 

i

 

 

K .

 

l l l l r l

1 2 '3 4 5 6
(length/width) ratio of ribbon

‘—

1N

further studies are necessary to understand the
mechanism completely. It is also necessary to test
scaled up components in order to investigate the ‘size
effect' before such composites can be incorporated into

structural applications.

QQEQLEEIQN§=

From the studies carried out so far it is
evident that discontinuous metallic-glass ribbon
reinforcements provide an unique method of obtaining
isotropic properties in brittle matrix composites.
These characteristics are as result of a combination of
factors which are 1) the high strength of the ribbons
2) the high interfacial bond strength, and 3) the
unique ribbon geometry. The stress transfer in such
composites is complex, and depends on both the ribbon
length and width. Specifying both these quantities is
essential, as they both affect the maximum load

carrying capabilities of the system.

 

The authors would like to thank the Composite
Materials and Structures Center at Michigan State

University for supporting and funding this project.

132

REFERENCES:

Vaidya R. U. and Subramanian K. N., J. Mater. Sci., 1;
(1990) 3291.

Vaidya R. U. and Subramanian K. N., J. Mater. Sci., (in
press).

Vaidya R. U. and Subramanian K. N., ASM/ACCE/ESD
Conference Proceedings, Detroit Oct. 89., published by
ASM, Metals Park OH., pp 227.

Vaidya R. U. and Subramanian K. N., J. Am. Ceram. Soc.,
(in press).

Vaidya R. U., Lee T. K. and Subramanian K. N., Fifth
annual American Society for Composites (ASC) Coference
Proceedings, East Lansing, June 90., published by
Technomic press, Lancaster PA., pp 902.

Strife J. R. and Prewo K. M., J. Mater. Sci., 11 (1982)
359.

Strife J. R. and Prewo K. M., AFWAL Report TR-80-4060
1980.

Vaidya R. U. and Subramanian K. N., manuscript

submitted to Comp. Sci. and Tech.

133

512511212

134

53513691:

The role of metallic-glass ribbons in modifying
the properties of glass-ceramics was investigated by
using specimens prepared by conventional pressing and
sintering techniques. Even very low volume fractions
of such reinforcements were found to provide
significant improvements in the strength, elastic
properties and fracture toughness of the glass-ceramic
matrices. The observed improvement in the fracture
toughness is explained on the basis of various
metallic-glass ribbon related energy absorbing

mechanisms.

135

IHIEQDQQIIQH=

Enhancement in the tensile strength and fracture
toughness of ceramics has been attempted by several
techniques such as microcrack toughening and
transformation toughening [1-4]. Recent studies have
shown that reinforcing ceramics with high strength
reinforcements is a viable alternative. In these
studies glass, conventional crystalline ceramics or
glass-ceramics are used as the matrices[5-7].
Potentials of various reinforcements including
continuous and discontinuous fibers(or whiskers) of
carbon, graphite, silicon carbide, alumina and various
metals like stainless steel and tungsten have been

investigated [5-11].

Among the various ceramic matrices, glass-
ceramics possess unique advantages. They are formed in
the glassy state and are converted to an almost 100%
crystalline state by subsequent heat treatment. Such a
feature facilitates low temperature composite
fabrication and at the same time provides for a
composite with high temperature capabilities (without
softening). Conventional crystalline ceramics(which

are formed by sintering powders) possess significant

136

137

porosity, which limits their strength. Glass-ceramics
on the other hand have little or no porosity, and hence

are tougher and stronger.

Potential of metallic glass ribbons as
reinforcements for ceramic matrix composites has not
been explored so far. Metallic glasses possess
superior fracture strengths and toughness as compared
to their crystalline counterparts. Metallic glasses
also possess good oxidation and corrosion resistance.
Their unique geometry provides for a large surface area
to bond with the matrix. Metallic glasses have been
studied as reinforcements for brittle polymer matrices
by Hornbogen et_§1 [12-14]. Significant improvement in
the mechanical properties of the polymer matrices were
reported by them. The main objective of the present
study was to develop metallic ribbon reinforced glass-
ceramic matrix composites and to evaluate their
mechanical properties. The nature of the metallic
glass/glass-ceramic interface and its role on the
mechanical properties of the composite system was also

of interest.

 

Two metallic glasses were used as reinforcements
in the present study; one was an iron-based metallic
glass METGLAS 26058-2 alloy, and the other a nickel-
based metallic glass METGLAS MBF-75 alloy. Both of
these metallic glasses were obtained from Metglas
Products, a buisness unit of Allied-Signal Inc. The
composition and properties of these metallic glasses as

provided by the manufacturer are listed in Table 1.

Based on initial experimentation and on the
basis of the low recrysatllization temperatures of the
two chosen metallic glasses, Corning glasses Code 7572
and 8463 were chosen as matrices. The compositions and
properties of both these glasses as provided by the

manufacturer are listed in Table 2.

Rectangular bar shaped specimens(6.25cm x 1.25am
x0.5cm) were made using the conventional wet pressing
and sintering techniques. Amyl acetate(3% of the
weight of the glass powders) was used as the binder.
After laying out the metallic glass ribbons

unidirectionally within the glass powders in a steel

138

139

TABLE 1: Properties of the metallic glass ribbons.

 

 

Property METGLAS 26058-2T METGLAS HEP-753T
Chemical composition : Fe : 78% Ni : 50%
B : 13% Co : 23%
Si : 9% Cr : 10%
Mo : 7%
Fe : 5%
B : 5%
CrystalIization temp. : 550°C 605°C
Elastic modulus : 85 GPa 70 GPa
Yield strength : > 700 MPa 1300 MPa
Coefficient of thermal _7 o _7 0
expansion : 76 x 10 / C 78 x 10 / C
*
Density : 7.18 g/cc 7.46 g/cc

T Code numbers of products of Metglas Products.
* Obtained from Ref. [22]. Rest of the entries
provided by the manufacturer.

 

140

TABLE 2: Properties of the ceramic glass matrices

(as specified by the manufacturer).

 

9 0

Property Corning Glass 7572 Corning Glass 8463

 

Softening point : 375°C 370°C

Coefficient of

thermal expansion : 95 x 10'7/ °c 105 x 10'7/ °c
Density (powder) : 3.8 g/cc 3.8 g/cc
(fired) : 6.0 g/cc 6.2 g/cc
Continuous service a a
temperature : 450 C 450 C
Chemical composition : PbO : 70% PbO : 84%
B 03 : 5-10% B 03 : 5-10%
5302 : 2-5: $102 : 2-5:
Al O : 1-5% Al 0 : 1-5%
2.8 : 10-20: 2n0 : 10-20:

9 code numbers of products of Corning Glass Co.

 

141

die, the composite specimens were pressed at 3000 psi.
After pressing, the specimens were first kept at 200°C
for 15 minutes to drive out the organic binder. The
specimens were then sintered at 400°C for 90 minutes.
Devitrification of the glassy matrix was carried out by
maintaining the composites at 450°C for 20 minutes.
After this treatment the specimens were furnace cooled
to room temperature so as to minimize the thermal

shock.

We:

The elastic properties of the unreinforced
matrix specimens and composite specimens were obtained
by the non-destructive Sonic Resonance Technique [15].
Since it was difficult—to detect the torsional
resonance frequency, the shear modulus was determined
by using the values of the Young's modulus (which was
obtained from the flexural resonant frequency), and by
assuming a Poisson's ratio of 0.25 for the composite

system.

Modulus of Rupture (MOR)measurements were made
by using the three point bend test in an Instron
testing machine with a cross head speed of 0.05cm/min.

The span to depth ratio for the specimens was

142

maintained in accordance with ASTM specification C-‘

203/85.

Static fracture toughness tests were performed
by fracturing Single Edge Notched Beam (SENB)
specimens, in three point bending. The notches were
cut using a diamond blade. The specimens were annealed
after cutting the notch, at 200°C, to heal up
microcracks which might have formed at the root of the
notch. The fracture toughness was determined using the
equations given by Gross and Srawley [16]. The
fracture toughness values for the unreinforced matrix
specimens were also determined by the non-destructive
indentation technique [17-18]. The specimen were
indented using a Vicker's indenter with a load of 0.3
kg. The fracture toughness was determined by using the
equations given by Lawn gt_§1_[17].

The pull-out test was carried out in order to
evaluate the interfacial bond strength. An embedded
length of 1.0 cm of ribbon(and width of 0.5cm) was used
for this purpose. In such a technique the interfacial
bond strength can be determined by balancing the
tensile forces to the shear forces acting on the

embedded portion of the ribbon.

M3
Fractographic studies were carried out using a

Scanning Electron Microscope.

The values of the elastic properties of the
unreinforced matrix and composite specimens as measured
by the Sonic Resonance Technique, are presented in
Table 3. A significant improvement in the elastic
properties is observed, even with the low volume
fraction of reinforcements used. The "Rule of
mixtures"(ROM) as used to characterize the elastic
properties of several composite systems is given by the
equation
V

EC 8 Eme + Ef f

where, ‘E' denotes the Young's modulus,‘V' the volume
fraction and the subscripts‘c',‘m' and‘f' refer to the
composite, matrix and ribbon respectively. The
calculated values of‘E' by using the ROM are compared
with the experimentally measured values, in Table 4.

As is evident from the results, the ROM does not

144

TABLE 3: Elastic properties of the glass-ceramic

matrices and composite systems obtained by the

Sonic Resonance Technique.

 

I
Class-Ceramic

 

Metallic-glass Volume E % increase
Matrix Reinforcement Fraction (GPa) in E
(COUfinngde) (METGLAS alloy) of reinforcement
7572 - 0% 33.4 -
7572 26053-2 0.73% 44.0 31.7
7572 26058-2 1.24% 47.7 42.8
7572 26058-2 1.648 69.4 108.0
7572 HEP-75 0.74% 42.1 25.9
8463 - 0% 28.1 -
8463 MBF-75 0.69% 36.0’ 28.0
8463 HEP-75 0.73% 40.8 45.4

 

M5

TABLE 4: Comparison of the experimentally measured and

calculated(by ROM) values of Young's modulus

for the METGLAS 26058-2 alloy reinforced 7572

glass-ceramic system.

 

Volume fraction

Young's modulus

Young's modulus

 

of reinforcement calculated by measured
(%) ROM experimentally
(GPa) (GPa)
0.73 33.78 44.03
1.24 34.04 47.70
1.64 34.24 69.43

 

146

characterize the elastic modulus of the composite
system under consideration. A better estimate of ‘E'
can be made by considering the equation given by Halpin

and Tsai [19-20], according to which,

E 1+n£Vf

 

E 1-17Vf

where, is the reinforcing efficiency, which will be
equal to one for a strongly bonded system, and
is an empirical constant which depends on
parameters like reinforcement aspect ratio, and

bond strength.

The value of ‘E ' can be obtained by fitting the
experimentally obtained values of ‘E' to the equation
given by Halpin and Tsai. For the system under
consideration the value of ‘ E ' is evaluated to be 0.24
(Figure 1). The value of the reinforcing efficiency
'11' was assumed to be unity since strong bonding was

observed between the ribbon and the matrix (Figure 2).

147

75 .
7031 .

 

25m

Young’s modulus E (GPa)

 

 

N—

o 1
Volume fraction of ribbons Vf (%)

FIGURE 1: Plot of the Young's modulus of the METGLAS
26058-2 alloy reinforced 7572 matrix
composites as a function of volume fraction

of metallic glass reinforcement.

FIGURE 2

(a

(b

V

v

 

Strong(void free) bonding between the METGLAS
26058-2 alloy ribbon and 7572 matrix.

The matrix is observed to be almost 100%
crystalline.

Matrix material adhering to the ribbon surface.

£11m:

The Modulus of Rupture (MOR) values of the
unreinforced matrix and the composite specimens are
presented in Table 5. In the system under
consideration, the reinforcing ribbons not only have a
higher fracture stress but also a higher fracture
strain as compared to the matrix. In the initial
stages of loading (in three point bending), the matrix
carries a major portion of the load. When the fracture
strength of the matrix is reached, the matrix cracks
and the load is transferred to the reinforcing ribbons.
Two different failure sequences can be enviasged
depending upon the volume fraction of reinforcements
used. For low volume fractions, when the matrix
cracks, the transfer of the load to the ribbons

overloads them and they fail. Hence,

ac = ome-l- of Vf
where, ac* is the fracture stress of the composite
°m* is the fracture stress of the matrix, and

of is the stress transferred to the

ribbons when the matrix cracks

M9

150

TABLE 5: Results of the three point bend tests carried

out on the glass-ceramic matrices and compos1te

specimens.

 

 

* value which does not agree w

Sonic Resonance Technique.

ith that obtained by the

Glass ceramic Metallic-glass Volume MOR % increase E
Matrix Reinforcement fraction of (MPa) in MOR (GPa)

( Corning Code ) (METGLAS alloy) reinforcement
7572 - 0% 14.98 - 26.15
7572 26058-2 0.80% 28.25 88.59 5.32*
7572 26058-2 1.24% 30.22 101.70 -
7572 26058-2 1.64% 41.25 175.40 -
7572 HEP-75 0.74% 32.27 115.39 -
7572 HEP-75 1.01% 33.25 121.96 -
8463 - 0% 11.30 - -
8463 MBF-75 0.68% 20.42 80.70 -
8463 MBF-75 0.69% 21.62 91.33 -
8463 HEP-75 0.71% 22.60 100.00 -
8463 HEP-75 0.73% 23.16 104.95 -
8463 MBF-75 0.77% 25.3 124.20 -

 

Ifl

When the volume fraction of the reinforcements
is high, the transfer of load to the ribbons is not
sufficient to fracture them and they continue to carry
the load until their fracture strength is reached.

Under these conditions

where, of* is the fracture stress of the ribbons.

The cross over point between these two
I
behaviours occurs at a critical volume fraction Vc ,

where

 

For the composite system under consideration the
calculated value of Va. is 0.5%. All the composite
specimens used in the current study had a volume
fraction of reinforcements greater than this critical

volume fraction. Hence the strength of the composite

um

-h~ 0‘
01 O
l A

404
35-
30- I
25- '
20d
15.:

Modulus of Rupture MOR (MPa)

 

 

Volume fraction of ribbons Vf (%)

l

FIGURE 3: Plot of the fracture strength of the METGLAS
26058-2 alloy reinforced 7572 matrix composites
as a function of volume fraction of metallic

glass reinforcement.

153

specimens is essentially a function of the volume
fraction of the metallic glass reinforcements. Higher
volume fractions of reinforcements should show
significant improvements in the fracture strength. The
variation in the MOR of the composite specimens with
increasing volume fraction of metallic glass

reinforcements is illustrated in Figure 3.

The Young's modulus (E) of the specimens was
also calculated from the results of the three point
bend test. The values of ‘E' obtained from the three
point bend test with those obtained from the Sonic
resonance test are compared in Table 6. The values of
‘E' obtained for the unreinforced matrix specimen by
both the techniques agree well; on the other hand the
values obtained for the composite specimen do not.
This discrepency may be attributed to the non-uniform

load carring characteristics of the composite system.

W:

The fracture toughness values for the
unreinforced matrix and composite specimens as measured
by the Single Edge Notched Beam technique, are listed
in Table 7. The fracture toughness values for the

unreinforced matrix specimens as measured by the

154

TABLE 6: Comparison of the values of the Young's modulus
of the METGLAS 26055-2 alloy reinforced 7572
glass-ceramic specimens obtained from the

Sonic Resonance and Three Point Bend tests.

 

 

Test Average Young’s Modulus Average Young's Modulus
7572 matrix 7572 + 26058-2 composite
(GPa) (GPa)
Dynamic 33.40 44.00
Resonance
Three point 26.15 5.32

bending

 

155

TABLE 7: Fracture toughness obtained by the Notched

Beam tests.

 

Sample KIC Average Standard COCff-Of
2

(maul/2) xIcmvamV ) Devaaml/Z) Variation

 

7572 matrix 0.4046
0.3580 0.378 0.0237 6.26%

0.3730

7572 matrix 1.0886
reinforced with 0.8320 0.952 0.3022 31.74%

0.6% METGLAS 1.1800

26058-2 0.7080

7572 matrix
reinforced with 1.372 1.401 0.041 2.93%
1.24% METGLAS 1.430

26058-2

 

156

Indentation technique, are listed in Table 8. The
indentation technique cannot be used to measure the‘
fracture toughness of the composite specimens because
the reinforcing ribbons are positioned far away from
the surface(where the indentation is carried out), and
as a result do not affect the crack growth behaviour at
the indentatation. A plot of the fracture toughness
versus the volume fraction of metallic glass
reinforcements is provided in Figure 4. A clear
enhancement in the fracture toughness with respect to
the unreinforced matrix is evident from this plot.

This behaviour can be attributed to improvements in the
various mechanical properties such as Young's modulus,
fracture stress and fracture strain, which can be
correlated to the fracture toughness using the
empirical equation given by Hahn and Rosenfield [21],

according to which

0.5
Ic “f ‘fL)
where, KIc is the fracture toughness
of is the fracture stress
at is the fracture strain, and

L is a geometrical correction factor.

In the system studied, the Young's modulus,

157

N
J

 

 

 

Fracture toughness K10 (MPa ‘V m )

j
i 2

Volume fraction of ribbons Vf (%)

0

FIGURE 4: Plot of the fracture toughness of the METGLAS
26058-2 alloy reinforced 7572 matrix composites
as a function of volume fraction of metallic

glass reinforcement.

158

TABLE 8: Fracture toughness obtained by the Indentation

technique.

 

Specimen Lead Borosilicate glass, code 7572.

Indentation load : 0.3 Kg.

Loading time 20 seconds.

50 micometers/sec.

Loading speed

Number of specimens : 3
Indentations
per specimen : 25
Fracture toughness : 0.496
1/2 0.433 Average : 0.46
x1c(upam ) 0.450

Standard deviation : 0.0327 urn-1’2

Coeff.of variation 7.11%

 

 

 

is"

159

fracture stress and fracture strain all increase with
increasing volume fraction of reinforcements, and hence
the fracture toughness is also expected to improve.

The improvement in the fracture toughness can also be
explained on the basis of fracture energy
considerations. The fracture toughness is related to
the Young's modulus and fracture energy (61c) by the

equation

0.5

K = (E GIG) .

Ic

The total energy absorbed during fracture of
the composite is the sum of the energies absorbed by
the matrix related processes (Gm) and by the ribbon
related processes (Gf).

Hence,
GIc 8 Gfo + vam.

The contribution to Gf arises from four
different ribbon related processes. In addition to
energy absorbed by ribbon failure (specific ribbon
fracture energy wf), energy is also dissipated as a
result of ribbon-matrix debonding (wd)and ribbon
pullout (wp). Furthermore there exists a bending
component (wb)as the crack in the surrounding matrix

opens before the reinforcing component is broken.

7...”.

1"

 

160

Hence,

It is believed that this crack deflection at the
ribbon-matrix interface, strengthens the composite and
makes it tougher by causing secondary processes such as
debonding and pullout to come into play, thereby
absorbing energy. This particular phenomenon can be
observed in Figure 5. Another mechanism of energy
absorption is the initiation of secondary cracks at the
edges of the reinforcing ribbons(Figure 6). These are
created when under the influence of bending moments,
the sharp edges of the ribbons tend to wedge open the
brittle matrix. The ribbons can also be assumed to
exhibit higher fracture strengths as a result of
hinderance of shear failure due to the surrounding
rigid matrix, increasing the contribution to wf

(Figure 7).

A very strong bond between the ribbon and matrix
was observed, as was evidenced by the absence of
ribbon-matrix debonding in the pull out test. Hence
the we and wp contributions are low in case of the
system under consideration. The main contribution to

the fracture energy of the present system are believed

1m

 

FIGURE 5: Crack arrest and deflection at the metallic

glass ribbon (METGLAS 26058-2)-matrix (7572)

162

 

‘FIGURE 6: Microcracks originating at the edges of the
reinforcing ribbons (METGLAS 26058-2) in the

7572 matrix.

163

matrix

 

FIGURE 7: Micrographs illustrating the high ductility

of the metallic glass ribbons.

(a) A crushed ribbon (METGLAS 2605S-2) in composite
failure.
(b) Vein type of fracture pattern on the metallic

glass (METGLAS 2605S-2) ribbon surface.
interface. The crack originated at the

tensile surface during the bend test.

164

to arise from the wf and “b components.
QQNQHQEIQH§=

1. Introduction of even a very low volume fraction of
metallic glass reinforcements, provide significant
improvements in the elastic properties, fracture
strength and fracture toughness of the brittle glass-
ceramic matrices. The strength of the composite system
is a function of the fracture strength and the volume
fraction of the ribbons, and increases proportionately

with increasing volume fraction of reinforcements.

2. The elastic properties of the present composite
system do not obey the rule of mixtures. They can be
predicted by using the empirical equation given by

Halpin and Tsai [19-20].

3. The improvement in the fracture toughness of the
present composite system is due to the introduction of
various ribbon related energy absorbing mechanisms such
as crack arrest and deflection and elastic bending and
fracture of the ribbons. Hicrocracking of the matrix at
the edges of the ribbons also contributes to the

fracture toughness.

The authors would like to acknowledge the

Composite Materials and Structures Center at Michigan
State University for supporting and funding this
project. The authors would also like to extend their
sincere thanks to Dr. Kenneth Chyung of Corning Glass
Works for providing the glass powders, and to Dr.
Edward Norin of Metglas Products for providing valuable

suggestions in the preparation of this manuscript.

165

REFE CBS:

10.

11.

K. Faber and A. Evans, J. Am. Ceram. Soc., 61
(1984) 255. l

A. Evans and A. Heuer, J. Am. Ceram. Soc.,é;
(1980) 241.

A. Evans, ”Advances in ceramics" Volume 12,
Edited by N. Claussen, M. Ruhle and A. Heuer,
publishers Am. Ceram. Soc., Ohio 1984, p 193.

R. McMeeking and A. Evans, J. Am. Ceram. Soc.,

la
01

(1982) 242.

R. Sambell, D. Brown and D. Phillips, J. Mat. Sci.,
(1972) 663.

R. Sambell, D. Brown and D. Phillips, 151g 1 (1972)
676.

M. Sahebkar, J. Schlichting and P. Schubert,
Berichte de Deutschen Keramischen Gesellschaft,

55 (1978) 265.

K. Gadkaree and K. Chyung, Am. Ceram. Soc. Bull.,§§
(1986) 370.

P. Shalek, J. Petrovik and F. Gac, Am. Ceram. Soc.
Bull., 95 (1986) 351.

S. Risbud and M. Herron, Am. Ceram. Soc. Bull., 65
(1986) 342.

A. Caputo, D. Stinton and T. Besmann, Am. Ceram.

166

12.

13.

14.

15.

16.

17.

18.

19

20.

21.

167

Soc. Bull., 66 (1987) 368.

A. Fels, K. Friedrich and E. Hornbogen, J. Mat. Sci.
Letters, 6 (1984) 569.

A. Fels, K. Friedrich and E. Hornbogen, J. Mat. Sci.
Letters, 3 (1984) 639. .

T. Tio, K. Friedrich, E. Hornbogen, U. Koster and
A. Fels, J. Mat. Sci. Letters, 3 (1984) 415.

S. Screiber, D. Anderson and N. Soga, ' s c
constants and their measurgments", publishers
McGraw Hill, New York, 1974, p.82.

M. Srinivasan and S. Seshadri, "Fracture megganics
for ceramicsI rggkg and Qggcrete", ASTM special
technical publication Nos. 745, 1977, p.46.

B. Lawn, " ra ur ch 3 of cer ' s" ,

Vol.5 Edited by R. Bradt, A. Evans and

D. P. H. Hasselmann, publishers Plenum press,

New York, 1983, p.1.

D. Shetty, A. Rosenfield and W. Duckworth, J. Am.

Ceram. Soc., 66 (1985) 65.

J. Halpin and S. Tsai, Air 26:96 Materials Lgpgratory

IQ_hniQQi.B§EQE§.AE!L:IB:§Z:AZ§. 1957-
D. Hull, ” t'o o s' ' ls"

Edited by R. Cahn, E. Davis and I. Ward, publishers
Cambridge University Press, Cambridge, 1985, p.127.
G. Hahn and A. Rosenfield, 2:991_3I§_IDL1_QQDI;_QD
Eracture, Mggioh, wee; Qggmany, 1 1973 . 3.

168

22. J. C. M. Li, "Treatise on Material Science and
Technolggy-Vol 20.", Edited by Herbert Herman,

publishers Academic Press, New York, 1981, p.325.

 

MICHIGAN smTE UNIV. LIBRARIES
lllllWW“111111101111ll1111111111111”Hl
31293007886785