NASA Contractor Report 180888 /d/BY3

1 ’
High Temperature Tensile Testing /

of Ceramic Composites
(EASA-CG-lSOSE&) BIGE TEHPEBAIUEE I E l S I L B li88-15S 56
IESILIYG GP C E b A I l C CCLtECSIlES Annual Report
( C l e v e l a n d S t a t e Cniv.) 46 F CSCL 14B
Unclas
G3/35 0 1; 1843

John 2. Gyekenyesi and John H. Hemann

Cleveland State University
Cleveland, Ohio

February 1988

Prepared for
Lewis Research Center
I Under Grant NAG3-749

National Aeronautics and

Space Administration
 HIGH TEMPERATURE TENSILE TESTING OF CERAMIC COMPOSITES
John 2. Gyekenyesi and John H. Hemann
Cleveland State University
Deparment of Civil Engineering
Cleveland, Ohio 44115

SUMMARY
The various components of a high temperature
tensile testing system are evaluated. The objective is
the high temperature tensile testing of Sic fiber
reinforced reaction bonded Si N specimens at test
temperatures up to 165OOC (30$08F). Testing is to be
conducted in inert gases and air. Gripping fixtures,
specimen configurations, furnaces, optical strain
measuring systems, and temperature measuring techniques
are reviewed. Advantages and disadvantages of the
various techniques are also noted.

INTRODUCTION
Ceramic composites are presently being developed
for high temperature use in heat engine and space power
system applications. These ceramic composites offer
increased toughness over monolithic ceramics as can be
seen by the idealized stress-strain curves shown in Fig.
1. The operating temperature range is expected to be up
to the limit of available fibers, perhaps eventually to
165OoC (3000OF) or above. Very little mechanical data is
available for fiber reinforced ceramic composites at
these high temperatures. The proper conduct of such
characterization tests requires the development of a
tensile testing system in preference to flexural tests.
Flexure tests produce maximum stresses only in a small
volume near the surface of a specimen while tensile
testing produces maximum stresses through the volume. A
tensile testing system includes unique gripping,
. heating, temperature and strain measuring devices which
require special considerations. The system also requires
an optimized specimen shape.
The purpose of this paper is to review various
systems and devices which may be applicable to the high
temperature tensile testing of unidirectional fiber
reinforced ceramic composites. The tests are to be
performed in air to determine their characteristics in
an environment which would be similar to that of the
composite's final application. The testing temperature
will be 165OOC (3000°F) which is approximately the use
limit of the available silicon based ceramics in an
oxidizing atmosphere. Some testing will be performed in
an inert atmosphere to eliminate the effects of
oxidation.
Until now, very few high temperature tensile
tests have been performed on composites in air, and
these tests have been limited to a maximum temperature
of 1000°C (1830OF) (refs. 1-7). The specimens in most of
these cases were Sic fiber reinforced glass-ceramic
matrix composites which have a lower operating
temperature range when used in structural components.
Our objective is to test Sic fiber reinforced reaction
bonded Si3N4 (SiC/RBSN) and Sic fiber reinforced Sic
(SiC/SiC) composites which can sustain higher
temperatures than the glass-ceramic matrix composites.
The various components of the tensile testing
system are covered individually. A number of grip
designs were evaluated to determine which techniques are
most applicable to the high temperature tensile testing
of longitudinal fiber reinforced ceramic composites.
The specimen size and shape were optimized based on a
number of limitations which will be discussed. This is
followed by an evaluation of various furnace or high
temperature chamber configurations. Finally, a number of
strain measuring techniques are described, followed by
some recommendations for the total tensile testing
system.
I. GRIPPING FIXTURES
For hilgh temperature tensile testing, it is
desired that the temperature of the grips should be the
same as the temperature of the specimen gage section.
This condition would minimize the effects of any
potential thermal gradients. Testing in an inert
atmosphere would allow the use of hot grips since
graphite can be utilized for the grip material. On the
other hand, testing in the presence of an oxidizing
atmosphere and at temperatures greater than 1000°C
(183O0F) results in a highly demanding application for
hot grips. At high temperatures, the grips may react
with the specimen, and they may also oxidize to cause
failure when testing in air. For high temperature
testing in air, most practical gripping fixtures require
some form of cooling. It is, however, still desirable to
have the grips as hot as possible to reduce thermal

2
gradients within the specimen. This is due to the fact
that such thermal gradients induce additional stresses.
In addition, unlike metals and monolithic ceramics, the
flat specimens of unidirectional fiber reinforced
ceramic composites are extremely weak in shear relative
to their longitudinal tensile strength. This limits the
types of grips which can be utilized.
For room temperature tensile tests, the use of
adhesively bonded tabs and serrated grips offers an
optimum setup. This technique is actually the
standardized method for testing polymer matrix
composites. For ceramic composites, this method gives
acceptable results but it cannot be applied at high
temperatures, due to the lack of an adhesive which
offers adequate strength at high temperatures.
Serrated grips have been applied to the high
temperature tensile testing of silicon carbide fiber
reinforced lithium aluminosilicate (SiC/LAS), but with
no success. The serrated grips were applied directly to
the surface of the specimen. As a result, the specimen
failed under the grips by shredding (ref. 3).
Another type of gripping fixture uses one or
more pins to apply the load to the specimen. The pins
are inserted through holes drilled in the specimen. The
holes have to be very accurate so that the load from the
individual pins is distributed as evenly as possible
across the thickness of the specimen. With the use of
more than one pin, the alignment of the holes becomes
very critical, since it is important to have the load
distributed as evenly as possible among the pins. High
accuracy of the hole alignment and concentricity are
very difficult to achieve in ceramic composite
specimens. This is due to the extreme hardness of
ceramics and the varying hardness of ceramic composites
as the different materials of the matrix and the fibers
are encountered during drilling. Some researchers have
experimented with pin grips at room temperature (ref. 8)
resulting in shear failure near the grips of the
specimen as illustrated in Fig. 2. Pin grips have been
applied successfully to cross-plied specimens consisting
of a ( 0 / 9 0 ) , layup (ref. 81, but it is difficult to
derive basic lamina characteristics from these tests.
Another compensating technique which has been applied
with the use of pin grips is the reduction of the
specimen gage section through the width and the
thickqess. This technique has promoted failure within
the specimen gage section, but the results show lower
tensile strengths relative to an as fabricated specimen

3
(ref. 8). Additional details of specimen configurations
are discussed later in this paper.
Friction grips have been applied to the tensile
testing of composites quite successfully. This includes
room temperature tests (ref. 9) as well as high
temperature tests (refs. 1-7). Most of the grips consist
of smooth faces with some type of insert such as Sic
paper or alumina to protect the specimen. The force can
be applied to the grips by pneumatic, hydraulic, or
mechanical means. The mechanical technique just uses
wedge grips. For testing at high temperatures the
material of the grips would most coveniently be a nickel
based alloy or silicon carbide. Researchers at General
Electric Company have run tensile tests with a grip
temperature of 98OoC (1800°F) (ref. 7) using a nickel-
base alloy as the grip material.
In tests conducted at the Massachusetts Institute
of Technology (ref. 3 ) , a single pin was used in
conjunction with the friction grips. The pin was used
for initially aligning the specimen. During the test,
the pin carried only a small portion of the total load
(ref. 3 ) . Another technique used for initial alignment
of the specimen with friction grips is to lightly load
the specimen and allow it to slip slightly within the
grips. The grips are then fully loaded and the specimen
is ready to be tested. At Southern Research Institute,
great care is taken to align the specimen within the
grips. At first, the circular specimen is loaded into
the upper grips. Next, the gripping fixture is rigidly
fixed such that it can be rotated about its axis. As it
is rotated, a dial indicator is used to check run-out at
the other end of the specimen. Then the lower grip is
clamped onto the other end of the specimen. Supported
vertically the straightness of the assembly is checked
at various points. Once the assembly is within
acceptable tolerances, it is loaded into the tensile
testing machine. This technique has been applied to the
testing of circular ceramic and carbodcarbon composite
specimens. With brittle materials, such as ceramics,
alignment is extremely critical (ref. 10) in order to
minimize unwanted bending stresses.
Another area of concern related to the gripping
fixture is the alignment device in the load train of
the tensile testing system. A number of alignment
devices are available for tensile testing equipment.
Most readily available equipment use universal joints
at the end of the load train with rigid grips. Universal
joints offer relatively low resistance to motion when
unloaded but the frictional forces increase

4
significantly as the applied load is increased. A more
sophisticated technique recently introduced by Oak Ridge
National Laboratories (ORNL) uses pistons and hydraulic
fluid in the alignment device (ref. 11). This is
illustrated in Fig. 3. The device consists of eight
pistons that are equally spaced in a circle with
hydraulic passages connecting their respective
cylinders. The pistons support the pull rod and the
hydraulic fluid distributes the loads between the
pistons such that all bending stresses are nearly
eliminated. The grips are part of the alignment device.
As a result, the specimen is located very close to the
alignment device which maximizes its effectiveness. This
system has been shown to produce bending stresses that
are less than 0.5% of the average applied tensile stress
(ref. 11).
A different technique utilizes air bearings at
the ends of the load train. This system has been applied
at Southern Research Institute. The system is basically
a ball and socket separated by high pressure air which
offers a virtually frictionless connection. A schematic
of the system is shown in Fig. 4 . Loads due to
extraneous crosshead motion are greatly reduced. In
addition, there are no rotational constraints (ref. 12).
Since the air bearings are located at the ends of the
load train, greater precision is required in loading
the specinem into the grips.

11. SPECIMEN CONFIGURATIONS

In the case of longitudinal fiber reinforced
ceramic composites, it is desirable to use a flat
straight sided specimen with tabs bonded to the ends.
This would result in a specimen with minimal stress
concentrations and minimal cost. Unfortunately, as noted
before, the tabs cannot be used due to the lack of an
adhesive which retains adequate strength at high
temperatures. Specimen size and availability also is
limited. At present, the length of the specimen is
limited to about 130 millimeters (5 inches) by cost and
available processing equipment. The short length of the
specimen dictates the use of fairly high temperature
grips. Otherwise, the thermal gradients would cause
significant stresses within the specimen. This requires
the ends of the specimen to be held at a high enough
temperature so that tabs cannot be used.
Most gripping fixtures induce stress
concentrations within the specimen which can cause the

5
specimen to fail near the grips outside of the gage
section. To promote failure within the gage section it
is necessary to neck down the specimen. Some researchers
have tested SiC/LAS composites which have been necked
down in both the width and the thickness directions
(refs. 1-5,8). The reduction of the specimen through the
thickness has been shown to reduce the strength of the
specimen. This is due to unavoidable damage to the
fibers of the outer lamina during machining of the
specimen (ref. 8). On the other hand, testing specimens
(SiC/LAS or SiC/CAS 11) which have reduced sections
through the width only have resulted in fairly
successful tensile tests (refs. 6 , 7 ) . Figure 5 shows a
SiC/LAS specimen with dimensions as used by researchers
at the University of Dayton Research Institute. This
specimen used a single radius for the reduced gage
section width (ref. 6). On the other hand, researchers
at General Electric used two radii for the reduced gage
section of the SiC/CAS I1 specimens (ref. 7).
For the testing of SiC/RBSN or SiC/SiC, it is
recommended that the specimen be reduced in the gage
section through the width and have a constant thickness
throughout (ref. 10). Some researchers also have
recommended this specimen configuration for SiC/LAS
composites (ref. 2). Caution, however, must be exercised
in the design of the specimen shape. A specimen which
has a reduced section through the width can fail by
fracturing at the radius with the crack propagating
longitudinally along the fibers up through the grip
section. The use of finite element analysis (FEA) is
recommended to optimize the specimen shape and increase
the likelihood of failure in the gage section. Selected
strength results from room temperature tests can be
utilized for validating the finite element analysis.

111. FURNACES AND HIGH TEMPERATURE CHAMBERS

Many techniques are available for raising the
temperature of the gage section of a tensile specimen
to the desired temperature of 165OoC (3000OF). In
addition to the temperature requirements, it is
necessary to have access to the specimen for such
instrumentation as extensometers or optical strain
measuring devices and some type of temperature measuring
gage. The furnace has to operate successfully for a
reasonable duration in an inert or an oxidizing
atmosphere. Also, due to the limited length of the
specimen, the furnace has to be generally small. For

6
 adequate representation of material strength
characteristics, it is desirable to have at least a 25
millimeter (1 inch) gage section with a homogeneous
c temperature distribution. Heating techniques include
resistance and induction furnaces, radiant furnaces,
laser heating, radio frequency induction heating, and
direct electrical heating (ref. 13). These techniques
are discussed in detail in the following paragraphs.
Indirect heating techniques, such as the
resistance, induction, and radiant furnaces are
relatively insensitive to the specimen material and
coatings. Their operating characteristics are
reproducible including the specimen temperature profile.
Disadvantages include possible ousceptor or heating
element material problems in air, less access to the
specimen gage section, and less efficient heating than
by direct means.
Direct heating techniques, such as laser heating,
radio frequency induction heating, and direct electrical
heating offer a relatively simple system and are
efficient techniques. The latter two techniques of
course require a conductive specimen. In addition,
direct heating requires minimal insulation and offers
good access to the specimen for strain and temperature
measuring devices. Disadvantages include power supply
matching problems since they are dependent on the
material of the specimen and it is difficult to
reproduce the testing conditions. It is also difficult
to control the specimen temperature profile.
Furnaces can be set up in one of two ways for
maintaining an inert atmosphere. One technique, used at
Southern Research Institute, employs an open furnace
which utilizes constant purging of an inert gas (ref.
10). This eliminates the complexities of sealing the
furnace but it also increases the turbulence within the
furnace. The other technique, which has been recommended
by other researchers (ref. 71, is to seal the furnace,
which minimizes the flow of gases and reduces the
turbulence at high temperatures. The reduction in
turbulence is required with the use of extensometers
with alumina rods or with optical strain and optical
temperature measuring devices. Researchers at General
Electric Company have two different furnaces: one for
testing in an inert atmosphere and another for testing
in air (ref. 7). Both furnaces are sealed, however, to
reduce the gas turbulence.

7
 1. RESISTANCE AND INDUCTION FURNACES
Resistance and induction furnaces are the most
common form of heating for general materials laboratory
work (ref. 13). They consist of electrical resistance
coils supported by a refractory molding. Resistance
furnaces have the heating elements in the same chamber
as the specimen. In contrast, induction furnaces have
the elements isolated from the chamber containing the
specimen. This allows the elements to operate in a
different atmosphere relative to the specimen. The coils
usually consist of multiple windings such that a number
of zones can be defined. The control of the individual
zones can be used to partly compensate for the heat loss
to the grips in a tensile testing system. A common and
convenient form of this furnace has the chamber divided
into half cylinders which are hinged together. This
allows for easy setup due to the ready access to the
grips and the specimen. The split furnace also allows
the use of windows for instrumentation such as
mechanical extensometers or various optical devices.
Different materials are available for use as
heating elements depending on the desired temperature
range and environment. Platinum alloys can be used up
to 180OOC (3300OF) in an inert or oxidizing atmosphere
(ref. 14) and some ceramics (e.g. silicon carbide) can
be used to well over 20OO0C (3600OF) (refs. 13-15), but
only in an oxidizing atmosphere. Molybdenum disilicide
(Super Kanthal) can be used to 175OoC (3180OF) in air
and 155OOC (2820OF) in argon (ref. 14). Researchers at
the Naval Research Laboratory have applied a
platinum-rhodium element furnace in the tensile testing
of ceramic composites. This permitted a gage section
temperature of 150OOC (2730OF) (ref. 16). Researchers at
General Electric Company have two different furnaces
using a Super Kanthal element furnace for testing in air
and a tungsten element furnace for testing in an inert
atmosphere (ref. 7). The tungsten elements can be used
up to a temperature of 255OoC (4620OF) in a vacuum or
inert atmosphere. Super Kanthal elements have been used
at the National Bureau of Standards in a split furnace
for testing at 13OO0C (2370OF) in air (ref. 17).
Temperature is typically varied by some form of
controller. The controller usually utilizes a
thermocouple placed near the specimen gage section to
monitor the temperature. Researchers have been able to
control the temperature to k0.5OC at 170OOC with the use
of thermocouples attached to the specimen gage section

8
(ref. 18). With the testing of ceramic specimens,
problems may be encountered in reaction of the
thermocouple with the specimen.
The resistance and induction furnaces are highly
developed and reliable devices. There are many
commercially available units which can be adapted to
most applications. Due to the limited length of the
SiC/RBSN composite specimens and the use of cooled grips
when testing in air, these furnaces would have to be
quite short with two zones at most. With a split
furnace, the windows or optical devices would have to be
realigned every time the furnace is opened up. A closed
furnace would make access to the specimen difficult but
realignment of the optics would not be required. When
considering the heating elements, platinum-rhodium or
Super Kanthal offer a quick response and a working range
that includes the desired temperature of 165OOC (3000OF)
in an oxidizing atmosphere. Tungsten elements can also
be used at the desired temperature but only in an inert
environment. In addition, tungsten elements also offer a
quick response. On the other hand, silicon carbide
elements offer a much slower response than metallic
elements requiring more time for heating or cooling.
Unfortunately, use of silicon carbide elements is
limited to an oxidizing atmosphere only. The oxygen is
needed to form a Si02 layer on the outer surface of the
element which prevents the diffusion of the silicon into
the atmosphere. The advantage of the silicon carbide
elements is that they last longer and can be used at
higher temperatures which can be a long term advantage.

2. RADIANT FURNACES
One type of radiant furnace uses high intensity
lamps with reflectors focused onto the specimen. The
furnace typically has four lamps equally spaced in a
circle about the specimen axis. The reflectors are
usually air or liquid cooled. As with resistance and
induction furnaces, radiant furnaces can be split with
the two halves being hinged together. This allows for
easy access to the specimen and the grips.
Another configuration uses multiple lamps placed
in close proximity to the specimen surface. The lamps
are arranged such that various heating zones exist
allowing for the minimization of thermal gradients along
the length of the specimen gage section.
Radiant furnaces use the same type of controller
as the resistance and induction furnaces. These

9
controllers use a thermocouple which is placed on or
near the surface of the specimen gage section.
Depending on the emissivity of the specimen
surface, the specimen can be heated to well over 1000°C
(1830OF). Higher temperatures can be attained with
specimens having emissivities approaching that of a
black body. Due to the single zone heating, the
reflector type of furnace requires a long specimen so
that a large enough gage section would have a
homogeneous temperature distribution. The multiple lamp
configuration offers multi-zone heating so that shorter
specimens can be utilized.
Some researchers have used radiant furnaces with
tungsten-halogen lamps for testing SiC/LAS composites up
to temperatures of 1000°C (1830OF) (refs. 4,6). Others
have tested monolithic ceramics up to temperatures of
160OoC (2910OF) (ref. 19) with the use of infrared
lamps.
The reflector type of radiant furnaces offer good
access to the specimen, although the radiation heat loss
from the specimen can be quite significant. In contrast,
the multiple lamp type of radiant furnaces offer minimal
radiation heat l o s s . Access to the specimen for various
instruments is limited with the multiple lamp
configuration.
Radiant furnaces offer a rapid response compared
to resistance and induction furnaces. This is due to
their minimal thermal inertia. The furnace can be
quickly cooled and reheated. This can be a considerable
advantage when a large number of specimens have to be
tested for statistical strength analysis. The trade off
is in the higher initial and maintenance costs of the
radiant furnaces.

3. LASER HEATING
A specimen can be heated by using a powerful
laser, such as a carbon dioxide (C02) laser. The C02
laser operates at a wavelength of 10.6 microns which is
in the infrared region of the electromagnetic wave
spectrum. The technique is best suited for planar
specimens. With the use of a beam splitter and mirrors,
both sides of a specimen can be heated. The technique
minimizes material problems but is sensitive to surface
reflectivity, which can be a problem as the specimen
oxidizes during a test. In addition, an unobstructed
line of sight is required for the laser. This technique
does produce a significant axial thermal gradient which

10
can be detrimental since it is important to have a
specimen gage section with a near homogeneous
temperature distribution (ref. 14).

4 . RADIO-FREQUENCY AND DIRECT ELECTRICAL HEATING

Radio-frequency heating is a simple technique

which heats a specimen by the use of electromagnetic
radiation. It consists of a liquid-cooled induction coil
with relatively few turns, with the specimen located at
the axis of the coils. The technique gives an
unobstructed view of the specimen, allowing the use of
mechanical or optical strain measuring and temperature
measuring devices. The heating can be highly localized
permitting the use of short specimens.
The matrix of the SiC/RBSN specimens is a poor
electrical and thermal conductor. As a result, the
radio-frequency heating technique is not recommended for
use with the SiC/RBSN composite specimens, since the
efficiency of the furnace is proportional to the
electrical and thermal conductivities of the specimen.
The technique would also produce significant axial
thermal gradiants due to the short length of the
specimen.
Direct electrical heating is a technique in
which large amounts of current are passed directly
through the specimen, permitting totally unobstructed
view of the specimen. Unfortunately as with the
radio-frequency heating method, this technique requires
a specimen with good electrical conductivity. In
addition, this technique induces significant axial
thermal gradiants which cannot be adequately controlled.
As a result, direct electrical heating cannot be
utilized. It should also be noted that these techniques
require either very high voltages or large amounts of
current. Consequently, these systems require very bulky
and costly equipment which can also be a significant
disadvantage relative to other heating techniques.

IV. STRAIN MEASURING TECHNIQUES

There are many techniques available for
measuring strain in a tensile specimen. These methods
include strain gages, the crosshead displacement of a
tensile testing machine, mechanical extensometers, or
various optical methods. Stated again, our initial
objective is the tensile testing of SiC/RBSN and SiC/SiC

11
composites at 165OoC (3000°F) in an ipert or an
oxidizing atmosphere. Due to the harsh environment in
which the specimens are to be tested, many of the
conventional strain measuring techniques cannot be used.
From room temperature tensile tests of silicon carbide
fiber reinforced silicon nitride (ref. 20) it has been
shown that for this material, first matrix cracking
occurs at approximately 0.1 percent strain. For the same
specimen, the ultimate strength occurs at approximately
0.5 to 0.6 percent strain. This is an approximate range
in which the strain measuring device has to operate.
Initially a brief description of the more
commonly used mechanical strain measuring techniques is
given. Major advantages and disadvantages of their
application to high temperature tensile testing of
ceramic composites are discussed. Next, a general
overview is given of various optical techniques and
their advantages and disadvantages noted (ref. 21).

1. CURRENTLY USED TECHNIQUES FOR MEASURING STRAIN

Currently, many of the strain measuring methods
applied to high temperature material characterization
tests are mechanical. These techniques include strain
gages, mechanical extensometers, and the crosshead
displacement measurement method.
Strain gages are well developed, widely used,
and offer low cost. In general, strain gages are simple
to install and offer adequate accuracy. Unfortunately,
strain gages are intolerant of elevated temperatures.
At present, the highest attained working temperature is
approximately 8OO0C (1470OF) (ref. 22). It is known
that, at elevated temperatures, problems with attachment
of the gage to the specimen and oxidation are
encountered. In addition, there are problems with the
connection of electrical leads. Thermal compensation is
also difficult, resulting in subsequent lower accuracy.
Mechanical extensometers can be applied at high
temperatures and offer adequate accuracy. This technique
usually requires divots or notches to be cut in the
specimen surface. The notches or divots may cause
significant stress concentrations in a specimen which
can initiate first matrix cracking. Recently, various
extensometer systems have been marketed which do not
require notches or divots in the specimen surface. These
systems use a light pressure applied to the
extensometers to hold them against the specimen surface.
Problems which may be encountered include slipping of

12
 the extensometers during a test (ref. 7 ) or breaking of
the rods when the specimen fractures (refs. 1,2). The
extensometers can also act as a heat sink which would
cause undesirable thermal gradients within the gage
section of the specimen. The materials used for the
extensometers also limit the performance. Quartz rods
are limited to approximately 1090°C (2000OF). On the
other hand, alumina rods can be used at the desired
temperature of 165OoC (3000OF) but alumina is very
'sensitiveto air circulation (ref. 6). The air
circulation about the rods causes different thermal
expansion rates at various points in the alumina,
producing bending in the rods, which induces a random
signal into the extensometer system. The induced noise
can be significant. As a result, the use of
extensometers with alumina rods requires heavy
insulation to minimize air circulation about the rods
(ref. 7 ) .
The crosshead displacement method is extremely
simple and can be used with most tensile testing
systems. It is also very inexpensive. The main
disadvantage is that it offers very low accuracy. With
the crosshead far removed from the specimen gage
section it is difficult if not impossible to account
for nonhomogeneous strain in a specimen. It is also
difficult to account for the strain in the load train.
Consequently, this method is not suited for applications
where high precision is required.

2. OPTICAL STRAIN MEASURING TECHNIQUES

Optical methods offer increased accuracy over
the crosshead displacement technique and do not have
the breakage problems of the mechanical extensometers.
There are a number of different types of optical
methods available for measuring strain or displacement.
These include laser interferometry, electrooptical
tracking, speckle photography, speckle interferometry,
and holography. In addition, photoelastic coatings and
conventional Moire methods can be used to measure
in-plane strain.
Difficulties can be encountered with the
application of an optical strain measuring technique to
high temperature tensile testing in an oxidizing or
inert atmosphere. Turbulence within the furnace causes
variations in atmospheric density which cause the index
of refraction of the gas to change. For most of the
optical methods the turbulence causes optical noise

13
which can be partly accounted for. The furnace window
can cause additional changes in diffraction which can
also be compensated for. If an open window is used,
then the flow of hot gases just increases the
turbulence. Therefore, one has to exercise caution so
that the resulting optical noise from the increased
turbulence does not interfere excessively with the
desired measurements.
Some of the above optical strain measuring
techniques usually pose serious problems when applied to
high temperature tensile testing. These include
photoelastic coating methods, conventional Moire
techniques and holography.
Photoelastic coating methods are simple, well
developed, and offer adequate accuracy. This technique
requires a coating to be placed directly on the
specimen surface. The coating is required to have a
similar Poisson's ratio to that of the composite
specimen and, in addition, it would have to be
maintained at high temperatures in an oxidizing
atmosphere. In general, the photoelastic methods are
limited to temperatures of only a few hundred degrees.
Conventional Moire techniques require a grating
to be placed directly on the specimen surface. As with
the photoelastic methods, conventional Moire techniques
are simple and well developed. Also like the
photoelastic methods, the Moire methods are generally
limited to lower temperatures, due to difficulties in
maintaining the grating. Although in one recent
experiment, the Moire technique has been applied at a
temperature of 137OOC (2500OF) (ref. 23), but this is
below the desired test temperature.
Holography is a truly noncontact optical
technique. It does not require any special preparation
of the surface of the specimen and it is also extremely
sensitive. For example, holographic interferometry
using a helium-neon laser has a sensitivity of
approximately 0.3 microns (12 microinches) (ref. 2 1 ) .
As a result, the system has to be isolated from any
external vibrations, making the installation or
setting-up of the system quite tedious. The system also
requires a laborious alignment of optics. Unfortunately,
holography is best suited to measure out-of-plane
displacements such as those usually encountered in
vibration analysis. As a result, most holography
applications are for flexural or transverse vibrational
analysis of a specimen.

14
 3 . REAL TIME TECHNIQUES

In this section a description of the various

techniques which provide real time output is given. Due
to availability of this real time output, all these
methods can be used in a closed loop tensile testing
system. Advantages and disadvantages of application to
the high temperature tensile testing of ceramic
composites are discussed for each method. The general
specifications such as typical gage lengths,
sensitivities, and ranges also are summarized.
It should be noted that these methods do not
offer a full field view of the strains or displacements
in a specimen; the output is just an average strain or
displacement over the gage length of the specimen. These
methods track two fiducial marks, with one at each end
of the gage section.
The fiducial marks used with the majority of the
real time techniques are flags which are either bonded
or mechanically fastened to the specimen. Due to the
harsh environment in which the specimen is to be tested,
serious problems with flags may arise. For example, the
adhesive used to bond the flags to the specimen may not
hold causing the flags to slip or even fall off the
specimen. This would invalidate all the strain data from
a test. On the other extreme, the flags may react with
the specimen causing stress concentrations which can
initiate matrix cracking in the immediate vicinity of
the flags. For the mechanically fastened flags slipping
may cause erratic data output. As with the bonded flags,
the mechanically fastened flags can also react with the
specimen to produce undesirable stress concentrations.

A . LASER INTERFEROMETRIC STRAIN GAGE

The laser interferometric strain gage (LISG)

(ref. 2 4 ) , also referred to as the Sharpe method,
consists of a single laser which illuminates two
fiducial markings on the specimen. The markings reflect
the laser light to either side of the incident beam
where each reflected beam projects a set of interference
fringes. The motion of the fringes is proportional to
the change in the relative displacement of the markings
on the specimen. Figure 6 shows a schematic of the LISG.
The change in the relative displacement between
the markings is given by (ref. 21):

15
Where: d = the relative displacement of the markings
ml and m2 = the fringe orders
), = the wavelength of the incident laser beam
a = the angle between the incident laser beam
and the reflected beams
The (Am +~m2)/2term in equation (1) gives the mean
value o$ the change in the fringe orders. This
averaging accounts for the rigid body motions. The
fringes can impinge on a screen and their movement
measured manually with respect to a fiducial mark, or
phototransistors or photoresistors can be used to sense
the motion of the fringes. With the use of a
minicomputer, the LISG has been applied to a closed
loop tensile testing system ( r e f . 21).
The markings consist of pyramidal shaped
indentations produced by a Vicker's diamond
micro-hardness indenter. These indentations can be
placed directly into the specimen surface or in the
surface of a tab which can be adhesively bonded to the
specimen. One of the requirements for the markings is
that the surface of the indentations must be
reflective.
The application of the Sharpe method to the
testing of ceramic composites would require the use of
adhesively bonded tabs. Ceramics are hard and brittle
making it difficult to place accurate Vicker's
indentations directly in the specimen surface. Also,
recent observations have shown that the surface of some
ceramics is translucent (ref. 21). As a result, the use
of indentations directly in the specimen surface would
not meet the requirements for a highly reflective
surface. In addition to the unique problems associated
with the use of ceramic specimens, the surrounding
environment will be an oxidizing atmosphere (air) at
165OoC (3000OF). This would typically require the use
of platinum tabs and a ceramic cement. The gage length
can be varied from 100 microns ( 3 . 9 mils) to about 800
microns (31 mils). The gage length is relatively small.
As a result, there is a high probability that first
matrix cracking may occur outside of the selected gage
section. This would prevent the strain gage from
registering the sudden increase in strain due to matrix
cracking. Another problem at high temperatures is that
the light emitted by the specimen can saturate the

16
photodiodes at the wavelength of the laser making the
LISG ineffective (ref. 24). This problem can be
overcome with the use of a more powerful laser and/or
the utilization of an optical filter.
Typical specifications for this method, when
used for ceramic testing, are shown in Table I. Also
shown in this table are the conditions under which the

Table I: LASER INTERFEROMETRIC STRAIN GAGE (LISG)

(ref. 21)

MAXIMUM APPLIED
TEMPERATURE - 140OOC (2550OF)
TARGET - PLATINUM TABS
GAGE LENGTH - 1 0 0 TO 800 MICRONS
(3.9 TO 31 MILS)
DISPLACEMENT
RANGE - 300 MICRONS ( 1 2 MILS)
RESOLUTION - 1 MICRON (39 MICROINCHES)
ERROR - 2 TO 3 PERCENT
technique has been applied. The maximum range of 300
microns (12 mils) is actually limited by the resolution
of the photodetectors. For values greater than 400
microns (16 mils) the fringes become too fine for the
photodetectors to resolve (ref. 21). Out-of-plane
displacements less than 25 microns (1 mil) do not cause
any significant errors (ref. 21). Convection currents
about the specimen have minimal effects due to the small
gage length. The LISG is not recommended for use with
the high temperature tensile testing of unidirectional
fiber reinforced ceramic composites due to the required
small gage length. As noted before, there is a high
probability that first matrix cracking can occur outside
of this gage section.

B. OPTICAL STRAIN ANALYZER

The optical strain analyzer as used by Southern
Research Institute consists of two telescopes mounted
in a "piggy-back" arrangement (ref. 12) as shown in

17
 Fig. 7. The system is an electro-mechanical-optical
design which tracks two flags on either end of the gage
length of a tensile specimen. The upper telescope tracks
the upper flags carrying with it the lower telescope.
The lower telescope at the same time tracks the lower
flag. Elongation is measured as the displacement of the
image in the lower telescope. Due to the relative
motion of the scopes, the rigid body motions do not
affect the displacement measurement from the gage
section.
Each flag is a pair of targets on opposite sides
on the specimen as shown in Fig. 8. Using a dual optical
system with an average output minimizes any
irregularities due to non-axial motion of the flags.
High intensity lights are shined through the slots in
the targets which are picked up by the telescopes. For
high temperatures where the targets become self
luminous, the high intensity lights are not required.
For each scope, the light passing through the
targets from both sides of the specimen are formed on a
rectangular aperture. The aperture contains small slits
which permit passage of the upper and lower edges of the
light beams. An optical chopper breaks up the light into
pulses with a fixed rate which a photodiode then
converts to an alternating voltage. The aperture and the
chopper are illustrated in Fig. 9. The light beams going
to the apertures are blocked alternately by the chopper.
When the scope is at the null point the output from the
photodiode, due to the light from the apertures, is
constant. This is shown in Fig. loa. When the flag moves
due to specimen deformation, the amount of light passing
through the apertures changes. As a result, the output
from the photodiodes varies as the light source from the
apertures alternates. This is illustrated in Fig. lob.
Using a closed loop system, the scopes are repositioned
to the point where the output voltage is the same. The
displacements of the scopes are recorded which
corresponds to the displacement of the specimen gage
section. For system power amplification, commercial high
fidelity audio amplifiers are used.
In high temperature tensile testing, this system
has been applied to the testing of carbodcarbon
composites in an inert atmosphere. The flags were made
of carbon and were clamped onto the specimen using
carbon fasteners. The design has been perfected to the
level where the fracture of the carbon specimen is not
affected (ref. 2 5 ) . In the case of a ceramic composite
specimen in an oxidizing atmosphere, the flags would
have to be redesigned. This includes selecting a

18
material and finding a way to attach the flags to the
specimen. As noted before, there is always the
possibility of the material of the flags reacting with
the specimen causing a stress concentration. On the
other hand, it is possible for the flags to slip on the
specimen making the data from the test invalid. In the
past, data was collected from hundreds of tensile tests
of carbon/carbon composites where only one side was
monitored and then where both sides were monitored. It
was found that both techniques resulted in the same
average strain but the range or percent error was
reduced significantly by monitoring both sides of a
specimen. Therefore, it is highly recommended that both
sides of the specimen should be monitored for strain
(ref. 10).
The optical strain analyzer allows a maximum
displacement of 6.35 millimeters (250 mils) without any
readjustment. It has a precision of 1.27 microns ( 5 0
microinches) (ref. 25). General specifications are
summarized in Table 11.
TABLE 11: OPTICAL STRAIN ANALYZER
(ref. 21)
MAXIMUM
TEMPERATURE - >2000°C (3600OF)
GAGE LENGTH 30 MILLIMETERS (1.2 INCHES)

DISPLACEMENT
RANGE - 6 . 4 MILLIMETERS ( 0 . 2 5 INCHES)
RESOLUTION - 1.3 MICRONS ( 5 1 MICROINCHES)
ACCURACY - 22.5 MICRONS (298 MICROINCHES)

Electrooptical tracking is a strain measuring

technique which is recommended by many researchers for
use with the high temperature tensile testing of ceramic
composites (refs. 8,261. The optical strain analyzer
offers optimum performance for use with high temperature
tensile testing. The optical strain analyzer is an old
system but it has been improved to the point where it is
highly reliable and offers good performance.

C. ZIMMER ELECTRO-OPTICAL EXTENSOMETER

The Zimmer system has been utilized by various

19
researchers at Pennsylvania State University, National
Bureau of Standards (NBS) (ref. 171, and the Government
Industrial Research Institute of Japan (ref. 27) for the
testing of ceramics. These tests included high
temperature tensile tests. The system consists of high
intensity lights which are shined past a set of flags
mounted on a tensile specimen. The flags block a portion
of the light and the resulting shadow is projected on an
array of photosensors. The position of the shadow on the
photosensors changes directly in proportion with the
change in distance between the flags when the specimen
is deformed. As a result, the output from the
photosensors is used to determine the change in relative
displacement of the flags.
The flags consist of alpha silicon carbide
(a-Sic) which are bonded to the specimen. The flags
sinter to the specimen surface during high temperature
testing (ref. 28). The flags are illustrated in Fig. 11.
In addition, researchers at NBS attached platinum wires
to each flag to enhance the visibility of the flags.

TABLE 111: ZIMMER ELECTRO-OPTICAL EXTENSOMETER

(refs. 17 and 28)
MAXIMUM
TEMPERATURE - -16OOOC (2910OF)
GAGE LENGTH - 25 MILLIMETERS ( 1 INCH)
RESOLUTION - 1 MICRON (39 MICROINCHES)
ACCURACY - -t2 MICRONS ( 2 7 9 MICROINCHES)

The Zimmer system has been applied with a

specimen gage section temperature of 130OoC (2370OF) in
air (ref. 28). Approximate values for resolution,
accuracy, and other relevent information for the Zimmer
system are shown in Table 111. The maximum test
temperature at which this system can be utilized is
approximately 160OoC (2910OF) with the use of an optical
filter (ref. 17).

D. ZYGO LASER DIMENSION SENSOR

The Zygo Laser Dimension Sensor is a relatively
new system. Consequently, it has not been applied to the
high temperature tensile testing of ceramic composites

20
yet, although various researchers at General Electric,
National Bureau of Standards, and Corning are presently
evaluating the system for this purpose. The system uses
a scanning laser which is shined past a pair of flags on
a tensile specimen and onto a receiver. The receiver
senses the edge of the shadow from the flags and with
the use of a processor the relative displacement of the
flags is determined.
Table IV contains selected data for the system.
With the use of an optical filter, the system can handle
specimen temperatures up to 1900°C (3450OF). The system
can also measure displacement up to 115 millimeters (4.5
inches). The accuracy is *0.0025 millimeters (-.0.0001
inches) (ref. 29).

Table IV: ZYGO LASER DIMENSION SENSOR

(ref. 29)

MAXIMUM
TEMPERATURE - 1900°C (3450OF)
GAGE LENGTH - <115 MILLIMETERS (4.5 INCHES)
ACCURACY - +0.0025 MILLIMETERS (-+O.OOO1
INCHES)

4. SPECKLE PHOTOGRAPHY AND SPECKLE INTERFEROMETRY

Speckle methods are truly non-contact techniques
that require no special surface preparation. These
techniques use laser illumination to form the speckles.
The speckles are actually random points of interference
of light reflected from various depths of a diffusely
reflecting surface. Speckle methods offer a full field
view allowing for the derivation of strains or
displacements at any point on the surface of the
specimen. However, these methods do not operate in real
time.
In speckle photography, a single divergent laser
is used to illuminate a diffusely reflecting surface
producing speckles. Next, a photograph is taken of the
surface before and after deformation of the object by
double exposure of the same film. The resulting

21
negative, called a specklegram, is used to derive
displacements at various points on the surface of the
specimen.
Speckle interferometry has a specimen
illuminated by two incident laser beams which are
symmetric about the normal to the surface. The two
incident beams interfere producing a grating on the
specimen surface. As with speckle photography, a
photograph is taken before and after deformation of the
specimen. The same film is used for each photograph;
therefore the film is double exposed. The superposition
of the gratings produces Moire fringes which are used
to derive displacements at various points on the
specimen surface.
Another form of speckle interferometry is one in
which the speckle pattern is sheared using a double
aperture camera in which a prism is placed in front of
one or both apertures. Speckle shearing interferometry
offers higher quality fringes than speckle photography
and conventional speckle interferometry. In addition, it
offers increased flexibility, and a relatively simple
setup compared with other speckle methods. A more
detailed description of the various speckle methods is
given in reference 21.
With ceramics, it is generally desirable to test
as many specimens as possible due to the statistical
nature of their failure strength. In addition, because
of the long cycle times of the various speckle methods
and the limited range these speckle methods offer, these
techniques are not recommended for the high temperature

TABLE V. (ref. 21)

SPECKLE PHOTOGRAPHY
DISPLACEMENT
RANGE - 36 MICRONS (1.4 MILS)
PRECISION - 0.36 MICRONS (14 MICROINCHES)

SPECKLE INTERFEROMETRY .

DISPLACEMENT
RANGE - 6 MICRONS (240 MICROINCHES)

PRECISION - 0.04 MICRONS (1.6 MICROINCHES)

22
tensile testing of ceramic composites. Some general data
for speckle photography and speckle interferometry are
shown in Table V. These speckle methods are overly
sensitive to out-of-plane motion and require a laborious
alignment of optics. Other problems which may be
encountered with the high temperature testing of ceramic
composites include changing of the surfaces of the
specimen due to oxidation during the test. Recent
observations have shown that the surface of some
ceramics are actually translucent which may cause
additional problems with speckle methods (ref.21).

V. TEMPERATURE MEASURING METHODS

The desired temperature of 165OOC (3000°F) for
the testing of ceramic composites limits the number of
temperature measuring techniques which can be applied.
Available methods include use of platinum-rhodium
thermocouples and various optical techniques. Many
researchers have used thermocouples in high temperature
testing in air (refs. 4,18,19,30,31), although some
difficulties can arise with the attachment of the
thermocouple to the surface of the specimen. Also,
attached thermocouples can cause thermal gradients
within the specimen.
Alternatives to thermocouples are various types
of optical pyrometers. These pyrometers determine the
temperature of a body from the energy that is emitted
by the specimen. As the temperature of a specimen is
increased, the total energy radiated from the surface
increases. This is illustrated in Fig. 12. The total
radiated energy is a function of the specimen's
temperature and surface condition. For an ideal radiator
or blackbody the total energy, integrated over all
wavelengths, is a function of temperature. This relation
is shown by equation (2) which is known as the
Stefan-Boltzmann law:
Eb = 0T4 (2)
where: Eb = the energy radiated per unit time and per
unit area b the ideal radiator
o = 5.669 x W/m2 K4 ( 0 . 1 7 1 4 x lo-*
Btu/h ft2 O R 4 ) Stefan-Boltzmann constant
T = absolute temperature
Unfortunately, most bodies are not ideal radiators. As
a result, most bodies radiate less energy than a perfect

23
blackbody. The energy emitted from a real body is
related to the energy radiated from an ideal blackbody
by the emissivity constant. This is shown by equation
(3):
E = €Eb (3)

where: E = the energy radiated per unit time and per

unit area by a non-ideal radiator
E: = emissivity

In a real body, the emissivity varies with temperature

and the wavelength of the radiation. A certain type of
non-ideal radiator, known as a gray body, has a constant
emissivity which does not change with the wavelength.
Figure 13 shows the emissive power as a function of
wavelength for the gray body and a real body relative to
a blackbody. The various optical pyrometers which are
available use this radiated energy to determine the
temperature of the specimen.
In general, radiation pyrometers collect the
total radiated energy covering the ultraviolet (W),
visible, and infrared (IR) wavelengths. The energy
source is assumed to act like a gray body with constant
emissivity and the emissivity is manually entered. The
units usually are not focused. These devices are
typically hand held units used for measuring
temperatures of kilns, furnaces, and oxidized metallic
surfaces. The response time is typically four to six
seconds. A common temperature range is 700° to 130OOC
(1290O to 2370°F), but with the use of calibrated
diaphragms, the temperature range can be extended to
20OO0C (3630OF) (ref. 32). Radiation pyrometers offer
limited accuracy due to the limited sensitivity of most
detectors. Most detectors are sensitive to only a
fraction of the total energy emitted by a body, and are
more sensitive at some wavelengths than at other
wavelengths (ref. 3 3 ) . The other source of error is the
assumed emissivity value. At high temperatures, this
device can easily be in error by several hundred degrees
in temperature.
IR pyrometers also are hand held units with
either focused or nonfocused optics. These devices
determine the temperature of the radiation source from
the IR region of the emitted energy. This type of
device can be used for temperatures within the range of
-50° to 420OoC (-58O to 7590OF). As with the radiation
pyrometers, this technique assumes the energy source
acts like a gray body with a constant emissivity as a

24
function of wavelength. The IR pyrometer is a narrow
band or monochromatic pyrometer in contrast with the
radiation pyrometer which is a wide band or broad range
pyrometer (ref. 32). The IR pyrometer also offers less
sensitivity than the radiation pyrometer but it offers
increased accuracy.
Ratio pyrometers measure the energy radiated at
two or more selected wavelengths. The technique assumes
that the emissivity does not change with temperature at
each of the selected wavelengths. With the appropriate
wavelengths selected, the ratio pyrometer offers
increased accuracy over other types of optical
pyrometers, especially if the emissivity of a body is
unknown (ref. 34). Typically the ratio pyrometer can be
used for measuring temperature from 800° to 250OoC
(1470O to 4530°F) (ref. 32). Oak Ridge National
Laboratories recently has utilized a ratio pyrometer to
measure and control the temperature of an alumina
specimen at approximately 1000°C (1830OF).
Finally, a commonly used optical pyrometer is
the disappearing filament type. This technique compares
the brightness of the radiation source with a known
source (filament). The technique is subject to error
due to human interpretation and eye fatigue, although
automatic models can minimize the errors (ref. 32).
The advantage of these optical pyrometers is that
they are truly non-contact methods which eliminate the
problems of attachment. Some researchers who are
presently using thermocouples intend to use an optical
pyrometer once other problems of high temperature
tensile testing have been overcome (ref. 6). Optical
pyrometers also have some drawbacks. Many of these
devices depend on a constant emissivity which may
actually change during a test. The change in emissivity
of a ceramic specimen can be due to oxidation of the
specimen surface during a high temperature test in air.
For some materials, the emissivity can change by an
order of magnitude (ref. 33). This change in emissivity
can lead to significant errors in determining the
temperature of the specimen gage section. Additional
input from the walls of the high temperature chamber and
the hot gases which are in the near or in immediate
proximity to the specimen may also be significant.

CONCLUSION
Many of the currently available components for a
high temperature tensile testing system have been

25
surveyed and evaluated. The tensile testing of
longitudinal fiber reinforced ceramic composite
specimens at temperatures up to 165OoC (3000OF) in air
results in a highly demanding application for many of
the components. Consequently, some compromises have to
be accepted with some of the components.
Due to the limitations imposed by the specimen's
characteristics, such as a relatively low shear strength
as compared to the longitudinal strength, it has been
determined that some form of friction grips would offer
optimum performance. The use of mechanical wedges may
offer the simplest and most cost effective solution for
the friction grips. A silicon carbide insert would have
to be utilized between the grips and the specimen. The
tests which are to be conducted in air would require the
use of cooled grips due to the lack of a grip material
which can withstand the test temperatures and offer
adequate strength without serious degradation. The
material of the grips would have to be a nickel based
alloy or monolithic Sic. On the other hand, when testing
in an inert atmoshere, hot grips may be utilized. The
material of the grips can be graphite, which offers
adequate strength at high temperatures in a nonoxidizing
environment.
The gripping fixtures require some form of
alignment device. The hydraulic unit used at ORNL is
recommended. Also, careful alignment of the specimen is
required when initially loading it into the grips.
Silicon carbide inserts should be placed between
the specimen and the grips to minimize possible failure
of the specimen under the grips. In addition, to promote
failure within the gage section, it is recommended that
the specimen be necked down through the width. Finite
element analysis should be utilized in optimizing the
specimen shape before conducting the actual high
temperature tensile tests.
To raise the temperature of the specimen, the use
of a resistance furnace is recommended. The furnace can
be the split type which makes available easy access to
the specimen and the grips. Two different furnaces
should be used; one for testing in air and the other for
testing in an inert atmosphere. Both furnaces would
require at least two windows, one on either side of the
furnace. These windows would be used for the optical
strain measuring device. Another window might be added
for use with an optical pyrometer. It is also
recommended that the furnaces be sealed to reduce the
turbulence within the chamber. The heating elements
should consist of Super Kanthal or silicon carbide for

26
the in air testing and tungsten elements for testing in
an inert atmosphere. The size of the furnaces is
determined by the specimen and the type of grips which
are to be utilized. The furnace which is to be used in
the oxidizing atmosphere would have to be sized to
accommodate a 130 millimeter (5 inch) specimen with a 25
millimeter (1 inch) gage section and cooled grips. The
grips are to be located outside of the furnace. The
furnace which is to be used in an inert atmosphere has
to accomodate a specimen with hot grips. The grips
should be located within the furnace.
The recommended technique for tracking strain is
the use of an electro-optical strain measuring device.
The Zimmer Electro-Optical Extensometer offers optimum
performance. This electro-optical strain measuring
device requires two windows located on diagnally
opposite sides of the furnace.
For the initial testing, it is recommended that
a platinum-rhodium thermocouple be used for measuring
the temperature of the specimen gage section. Once
other problems have been minimized or eliminated the
use of an optical pyrometer is suggested. The use of
optical pyrometers requires some knowledge of the
specimen emissivity at the test temperatures. As a
result, determining the emissivity and how it changes
during a test can be very time consuming. Once the
emissivity of the specimen has been determined at the
test temperatures, the application of the ratio
pyrometer is recommended.

REFERENCES
1. M a n d e l l , J . F . , Grande, D.H., and D a n n e m a n n , K.A.,
"High Temperature Testing of Glass/Ceramic Matrix
Composites," ASTM Symposium, "Test Methods and
Design Allowables for Fibrous Composites: Second
Symposium," Phoenix, Az., Nov. 3 - 4 , 1986.
2. Mandell, J.F., Grande, D.H., and Jacobs, J.,
"Tensile Behavoir of Glass/Ceramic Composite
Materials at Elevated Temperatures," Gas Turbine
Conference and Exhibit, May 31-June 4 , 1987, ASME
paper no. 87GT75.
3 . Grande, D.H., "Testing and Properties of High
Temperature Glass-Ceramic Matrix Composites,11
Ph.D. Thesis, Massachusetts Institute of
Technology, 1987.

27
 4 . Mah, T:, et. al., "High Temperature Mechanical
Behavoir of Fiber Reinforced Glass Ceramic Matrix
Composites," Communication of the American Ceramic
Society, Sep 1985.
5. Mah, T., et. al., "Recent Developments in Fiber
Reinforced High Temperature Ceramic Composites,"
Ceramic Bulletin, V66, N2, Feb 1987, pp. 304-308.
6. Personal communication with Noel E. Ashbaugh,
University of Dayton Research Institute, Automated
Material Characterization Group, Jun. 5, 1987.
7. Personal communication with David A. Utah, General
Electric Company, Cincinnati, OH, August 21, 1987.
8. Larsen, D.C., et. al., "Test Methodology for
Ceramic Fiber Composites: Results for Sic/LAS,
SiC/SiC, and C/SiC Composites," In NASA, Langley
Research Center Metal Matrix, Carbon, and Ceramic
Matrix Composites 1985.
9. Jamet, J.F., Lewis, D., and Luh, E. Y.,
"Characterization of Mechanical Behavoir and
Fractogaphic Observations on Compglass SiC/LAS
Composites,'' Ceramic Engineerinq and Science
Proceedinqs, V5, Jul-Aug, 1984, pp. 625-642
10. Personal communication with H.S. Starrett, Southern
Research Institute, Birmingham, Al., Aug. 5, 1987.
11. Lui, K.C., and Brinkman, C.R., "Tensile Cyclic
Fatigue of Structural Ceramics," Proceedings of
the Twenty-Third Automotive Technology Development
Contractors' Coordination Meeting," Dearborn, Mi.,
Oct 21 to 24, 1985.
12. Starrett, H.S., "High Temperature Tensile Testing
in Air," Inter Agency Planning Group Meeting at
the Institute for Defense Analysis, Alexandria,
Virginia, Nov 12, 1986.
13. Sumner, G.,"Heating Methods and Grips," Techniques
for High Temperature Fatigue Testinq. London and
New York: Elsevier Applied Science Publishers,
1985, pp. 71-96.
14. Huddleston, R.,"Oak Ridge Presentation," Hiqh

28
 Temperature Testinq of Carbon-Carbon and Ceramic
Matrix Composites., ed. T. F. Kearns. Institute
for Defense Analyses, IDA Memorandum Report M-278,
Dec 1986.
15. Halbach, C.R., Page, R.J., Arthur, P.D., "2200°C
Oxidizing Atmosphere Furnace for Space
Manufacturing,'' AIAA 12th Aerospace Sciences
Meeting, Washington, D.C., Jan 30 to Feb 1, 1974,
AIAA Paper No. 74-154.
16. Lewis, D.! Bulik, C., and Shadwell, D.,
"Standardized Testing of Refractory Matrix/Ceramic
Fiber Composites," Ceramic Engineerinq and Science
Proceedinqs, V6, Jul-Aug, 1985, pp. 507-517.
17. Personal communication with Dan Carroll, National
Bureau of Standards, Gaithersburg, Maryland,
September 1, 1987.
18. Davies, C.K.L., and Sinha Ray, S.K.,"A Simple
Apparatus for Carrying Out Tensile Creep Tests On
Brittle Materials Up To Temperatures of 175OoC,"
Journal of Physics E: Scientific Instruments, V4,
1971.
19. Matsusue, K., Fujisawa, Y., and Takahara, K.,
"Tensile Test of Pressureless Sintered Silicon
Nitride at Elevated Temperatures," NASA Technical
Memorandum 77777.
20. Bhatt, R.T.,"Mechanical Properties of Sic
Fiber-Reinforced Reaction-Bonded Si3N4 Composites,"
NASA Technical Memorandum 87085.
21. Gyekenyesi, J.Z., and Hemann, J.H.,"Optical Strain
Measuring Techniques for High Temperature Tensile
Testing," NASA Contractor Report 179637.
22. Hobart, H.F., "The NASA Lewis Strain Gauge
Laboratory - An Update," Turbine Engine Hot
Section Technology 1986, NASA Conference
Publication 2444.
23. Bayer, M., and Cloud, G.,"Moire to 137OoC,"
Proceedings of the 1986 SEM Spring Conference on
Experimental Mechanics, New Orleans, LA, Jun. 8-13,
1986.

29
24. Jenkins, M.G.,"Ceramic Crack Growth Resistance
Determination Utilizing Laser Interferometry,"
Ph.D. Thesis, The University of Washington, 1987.
25. Pears, C.D., et. al., "Test Methods for High
Temperature Materials Characterization," Technical
Report AFML-TR-79-4002
26. Mecholsky, J.J.,"Evaluation of Mechanical Property
Testing Methods for Ceramic Matrix Composites,"
Ceramic Bulletin, V65, N2, 1986.
27. Wakai, F., et. al., "Tensile Creep Test of Hot
Pressed Si3N4,I1 Presented at the International
Symposium on Ceramic Components for Engines,
Hakone, Japan, Oct 17-21, 1983.
28. Wiederhorn, S.M., et. al., "Standard Tensile Test
Development," Proceedings of the Twenty-Forth
Automotive Technology Development Contractors'
Coordination Meeting P-197, Dearborn, Michagan, Oct
27-30, 1986.
29. "1100 Series Measurement and Control System/ Laser
Dimension Sensor Model 1104", Zygo Corp.,
Middlefield, CT 06455-0448.
30. Govila, R.K.,"High Temperature Uniaxial Tensile
Stress Rupture Strength of Sintered Alpha Sic,"
Journal of Materials Science, V18, 1983.
31. Rozak, G.A.,"Stress Rupture Behavoir of Silicon
Carbide Coated, Low Modulus Carbodcarbon
Composite," Master of Science Thesis, Case Western
Reserve University, 1987.
32. Jaroszynski, W.S., "Measuring Temperatures
Radiometrically,I' Machine Desiqn, July 9, 1987.
33. "Radiation Pyrometry, Theory and Practice,"
Capintec Instruments Inc., Montvale, NJ 138630.
34. Tenney 111, A.S., "Red Hot....and Hotter,"
Mechanical Engineering, V108, N10, Oct 1986.

30
 I
yULTIMATE ST7.E:; 3 TH

TENSILE CERAMIC COMPOSITE

STEf$S BEl+AVIO R

TENSILE STRAIN

Fig. 1 Idealized stress-strain curve for fiber

reinforced ceramic composites.

31
 w
0 0
0

Fig. 2 Shear failure of longitudinal fiber reinforced

ceramic specimen using pin grips.

32
 1. hydraulic piston assembly

2. fluid channel
3. fill port

4. oil chamber

5. O-ring seal io
11
L
6. piston
12
7. ball-bearing bushing
13
8. snap ring

9. cover 14

10. metal pull rod 15

11. fastener 16

12. grooved flat ring

13. cooling channel

14. tapered Split collar

15. metal plug

16. specimen

Fig. 3 Cutaway view of ORNL self-alligning Qrip system.

33
 I Positiaring Crosshead
I
r e c i s i o n madcell

as Exit

Cas P r e s s u r e

upper S p n e r i c a l
Cas Bearing
I
St8 Balance tlcatcr

Furnace

Specimen

Lover Spherical
Cab Dcrrring
- be Balance Heater
Precision G u i c k s

I I
I Power C r e s s i r e a d

Fig. 4 Schematic arrangement of gas bearing

universals, specimen, and load train.

34
 -r
4.00
P t 2 .so

ALL DIMENSIONS IN INCHES

Fig. 5 Ceramic composite specimen configuration

for high temperature tensile testing with a
friction gripping system.

35
 PHOTO DIODE

FRINGE PATERN

LASER

............
... ......
..... ...- ........
............
. . . . - . . . . . - I

.............
............
L

.............
............
............
............
...
..
.:.-$:p Ecri;i &J
...........
...........
...........
...........
........ -.
,..........
..........
..........
,........
..........
.........
,.........
........
.......
.......
’.
......
......
.....
’.
.....
’.

Ad =
Am, + Am2 - x
2 SIN a

Fig. 6 Schematic of the Laser Interferometric Strain

Gage.
36
l -

37
 TENSILE SPECIMEN

A X I A L STRAIN
ATTACHMENTS

&
I
FRONT VIEW t

EDGE VIEW

Fig. 8 Targets attached to tensile specimen for use

with the Optical Strain Analyzer.

I 38
w
a
3
c
a
nl
0
a
Y
4 L
0,
i
V
(L

-Q!
C
VI
v)

a
V

39
 W

TIME

T T

TIME

Fig. 10 Output for the Optical Strain Analyzer:

a) Null point before deformation of the
specimen.
b) After deformation of the specimen.

40
 CIMEN

E-
TOP VIEW

l
il
FRONT VIEW. SIDE VIEW
FLAG
.

Fig. 1 1 Flags attached to tensile specimen for use

with the Zimmer Electro-Optical
Extensometer.

41
 TEMPERATURE, T
T, > Tr Tj

.
WAVEL ENG T H

Fig. 12 Emissive power of a body as a function of

wavelength and temperature.
~ 42
 € A = EMISSiVITY AS A FUNCTION OF WAVELENGTH
TEMPERATURE T = CONSTANT

GRAY BODY
4 CONST,

WAVELENGTH ,X

Fig. 13 Emissive power of various bodies as a

function of wavelength.
43
 NaliOnal Aeronaulics and Report Documentation Page
1. Report No. 2. Government Accession No. 3. Recipieiit's Catalog No.
NASA CR-180888
4. Title and Subtitle 5. Report Date
February 1988
High Temperature T e n s i l e T e s t i n g o f Ceramic Composites

7. Author@) 8. Performing Organization Report No.

John Z. Gyekenyesi and John H. Hemann None
10. Work Unit No.

- -I
506-43-1 1
9. Performing Organization Name and Address
Cleveland State U n i v e r s i t y
Deeartment o f C i v i l E n g i n e e r i n g
Cleveland, Ohio 44115-
I 11. Contract or Grant No.

NAG3-749
13. Type of Report and Period Covered
12. Sponsoring Agency Name and Address C o n t r a c t o r Report
Annual
N a t i o n a l A e r o n a u t i c s and Space A d m i n i s t r a t i o n 14. Sponsoring Agency Code

Lewis Research Center

C1 eve 1 a n d , Ohio 441 35-31 91

16. Abstract

The v a r i o u s components o f a h i g h temperature t e n s i l e t e s t i n g system a r e e v a l -

uated. The o b j e c t i v e i s t h e h i g h temperature t e n s i l e t e s t i n g o f Sic f i b e r
r e i n f o r c e d r e a c t i o n bonded Si3N4 specimens a t t e s t temperatures up t o 1650 "C
(3000 O F ) . T e s t i n g i s t o be conducted i n i n e r t gases and a i r . G r i p p i n g f i x -
t u r e s , spec men c o n f i g u r a t i o n s , f u r n a c e s , o p t i c a l s t r a i n measuring systems, and
temperature measuring techniques a r e reviewed. Advantages and disadvantages o f
t h e v a r i o u s techniques a r e a l s o noted.

7. Key Words (Suggested by Author@)) 18. Distribution Statement

High temperature t e s t Unclassified - Unlimited

O p t i c a l s t r a i n gauge S u b j e c t Category 35
Tensile t e s t i n g
Ceramic m a t r i x composites
9. Security Classif. (of this report) 20. Security Classif. (of this page) 21. No of pages 22. Price'
Unc 1 as s if 1 e.d Uncl a s s i f ied 45 A03