ASM Metals HandBook Vol. 8 - Mechanical Testing and Evaluation

Подождите немного. Документ загружается.

Significance and Use of Tensile-Testing Data for Elastomers. It is important to note that the tensile properties

of elastomers are determined by a single application of progressive strain to a previously unstressed specimen

to the point of rupture, which results in a stress-strain curve of some particular shape. The degree of

nonlinearity and in fact complexity of that curve will vary substantially from compound to compound. Tensile

properties of elastomers also have different significance than those of structural materials.

Tensile Strength of Elastomers. Because elastomers as a class of materials contain a substantial number of

different polymers, the tensile strength of elastomers can range from as low as 3.5 MPa (500 psi) to as high as

55.2 MPa (8.0 ksi); however, the tensile strengths of the great majority of common elastomers tend to fall in the

range from 6.9 to 20.7 MPa (1.0–3.0 ksi).

It should also be noted that successive strains to points just short of rupture for any given compound will yield a

series of progressively different stress-strain curves; therefore, the tensile-strength rating of a compound would

certainly change depending on how it was flexed prior to final fracture. Thus, the real meaning of elastomer

tensile strength may be open to some question. However, some minimum level of tensile strength is often used

as a criterion of basic compound quality, because the excessive use of inexpensive ingredients to fill out a

formulation and lower the cost of the compound will dilute the polymer to the point that tensile strength

decreases noticeably.

The meaning of tensile strength of elastomers must not be confused with the meaning of tensile strength of

other materials, such as metals. Whereas tensile strength of a metal may be validly and directly used for a

variety of design purposes, this is not true for tensile strength of elastomers. As stated early in ASTM D 412,

“Tensile properties may or may not be directly related to the end-use performance of the product because of the

wide range of performance requirements in actual use.” In fact, very seldom if ever can a given high level of

tensile strength of a compound be used as evidence that the compound is fit for some particular application.

Elongation of Elastomers. Ultimate elongation is the property that defines elastomeric materials. Any material

that can be reversibly elongated to twice its unstressed length falls within the formal ASTM definition of an

elastomer. The upper end of the range for rubber compounds is about 800%, and although the lower end is

supposed to be 100% (a 100% increase of the unstressed reference dimension), some special compounds with

limits that fall slightly below 100% elongation still are accepted as elastomers.

Just as with tensile strength, certain minimum levels of ultimate elongation are often called out in specifications

for elastomers. The particular elongation required will relate to the type of polymer being used and the stiffness

of the compound. For example, a comparatively hard (80 durometer) fluoroelastomer might have a requirement

of only 125% elongation, whereas a soft (30 durometer) natural rubber might have a minimum required

elongation of at least 400%.

However, ultimate elongation still does not provide a precise indication of serviceability, because service

conditions normally do not require the rubber to stretch to any significant fraction of its ultimate elongative

capacity. Nonetheless, elongation is a key material selection factor that is more applicable as an end-use

criterion for elastomers than is tensile strength.

Modulus of the Compound. Another characteristic of interest is referred to in the rubber industry as the

modulus of the compound. Specific designations such as 100% modulus or 300% modulus are used. This is due

to the fact that the number generated is not an engineering modulus in the normal sense of the term, but, rather,

is the stress required to obtain a given strain. Therefore, the 100% modulus, also referred to as M-100, is simply

the stress required to elongate the rubber to twice its reference length.

Tensile modulus, better described as the stress required to achieve a defined strain, is a measurement of the

stiffness of a compound. When the stress-strain curve of an elastomer is drawn, it can be seen that the tensile

modulus is actually a secant modulus—that is, a line drawn from the origin of the graph straight to the point of

the specific strain. However, an engineer needing to understand the forces that will be required to deform the

elastomer in a small region about that strain would be better off drawing a line tangent to the curve at the

specific level of strain and using the slope of that line to determine the approximate ratio of stress to strain in

that region. This technique can be utilized in regard to actual elastomeric components as well as lab specimens.

Tension Set. A final characteristic that can be measured but that is used less often than the other three is called

tension set. Often, when an elastomer or rubber is stretched to final rupture, the recovery in length of the two

sections resulting from the break is less than complete. It is possible to measure the total length of the original

reference dimension and calculate how much longer the total length of the two separate sections is. This is

expressed as a percentage. Some elastomers will exhibit almost total recovery, whereas others may display

tension set as high as 10% or more. Tension set may also be measured on specimens stretched to less than

breaking elongation.

The property of tension set is used as a rough measurement of the tolerance of high strain of the compound.

This property is not tested very often, but, for some particular applications, such a test is considered useful. It

could also be used as a quality-control measure or compound development tool, but most of the types of

changes it will detect in a compound will also show up in tests of tensile strength, elongation, and other

properties, and so its use remains infrequent.

Tensile-Test Curves. Figure 24 is a plot of tensile-test curves from five very different compounds, covering a

range of base polymer types and hardnesses. The contrasts in properties are clearly visible, such as the high

elongation (>700%) of the soft natural rubber compound compared with the much lower (about 275%)

elongation of a soft fluorosilicone compound. Tensile strengths as low as 2.4 MPa (350 psi) and as high as

15.5% MPa (2.25 ksi) are observed. Different shapes in the curves can be seen, most noticeably in the

pronounced curvature of the natural rubber compound.

Fig. 24 Tensile-test curves for five different elastomer compounds

Figure 25 demonstrates that, even within a single elastomer type, contrasting tensile-property responses will

exist. All four of the compounds tested were based on polychloroprene, covering a reasonably broad range of

hardnesses, 40 to 70 Shore A durometer. Contrasts are again seen, but more in elongation levels than in final

tensile strength. Two of the compounds are at the same durometer level and still display a noticeable difference

between their respective stress-strain curves. This shows how the use of differing ingredients in similar

formulas can result in some properties being the same or nearly the same whereas others vary substantially.

Fig. 25 Tensile-test curves for four polychloroprene compounds

Tests for Determining the Tensile Strength of Fibers

Mechanical properties of fibers are very dependent on test method. Two basic methods are the single-filament

tension test and the tow tensile test of a group or strand of fibers.

Single-filament tensile strength (ASTM D 3379) is determined using a random selection of single filaments

made from the material to be tested. Filaments are centerline-mounted on special slotted tabs. The tabs are

gripped so that the test specimen is aligned axially in the jaws of a constant-speed movable-crosshead test

machine. The filaments are then stressed to failure at a constant strain rate. For this test method, filament cross-

sectional areas are determined by planimeter measurements of a representative number of filament cross

sections as displayed on highly magnified micrographs. Alternative methods of area determination use optical

gages, an image-splitting microscope, a linear weight-density method, and others.

Tensile strength and Young's modulus of elasticity are calculated from the load elongation records and the

cross-sectional area measurements.

The specimen setup is shown in Fig. 26. Note that a system compliance adjustment may be necessary for

single-filament tensile modulus.

Fig. 26 Schematic showing typical specimen-mounting method for the single-filament fiber tension test

(ASTM D 3379)

Tow Tensile Test (ASTM D 4018). The strength of fibers is rarely determined by testing single filaments and

obtaining a numerical average of their strength values. Usually, a bundle or yarn of such fibers is impregnated

with a polymer and loaded to failure. The average fiber strength is then defined by the maximum load divided

by the cross-sectional area of the fibers alone.

Using ASTM D 4018 or an equivalent is recommended. This is summarized as finding the tensile properties of

continuous filament carbon and graphite yarns, strands, rovings, and tows by the tensile loading to failure of the

resin-impregnated fiber forms. This technique loses accuracy as the filament count increases. Strain and

Young's modulus are measured by extensometer.

The purpose of using impregnating resin is to provide the fiber forms, when cured, with enough mechanical

strength to produce a rigid test specimen capable of sustaining uniform loading of the individual filaments in

the specimen.

To minimize the effect of the impregnating resin on the tensile properties of the fiber forms, the resin should be

compatible with the fiber, the resin content in the cured specimen should be limited to the minimum amount

required to produce a useful test specimen, the individual filaments of the fiber forms should be well

collimated, and the strain capability of the resin should be significantly greater than the strain capability of the

filaments.

ASTM D 4018 Method I test specimens require a special cast-resin end tab and grip design to prevent grip

slippage under high loads. Alternative methods of specimen mounting to end tabs are acceptable, provided that

test specimens maintain axial alignment on the test machine centerline and that they do not slip in the grips at

high loads. ASTM D 4018 Method II test specimens require no special gripping mechanisms. Standard rubber-

faced jaws should be adequate.

Mechanical Testing of Polymers and Ceramics

Mechanical Testing of Ceramics

Ceramic materials have been used in a variety of engineering applications that utilize their wear resistance,

refractoriness, hardness, and high compression strength. Traditionally, they have not been used in tensile-

loaded structures because they are brittle and experience catastrophic failure before permanent deformation.

Nevertheless, their extreme refractoriness, chemical inertness, and favorable optical, electrical, and thermal

properties are inducements to use ceramics in certain tensile load-bearing applications. Typical mechanical

properties of common ceramics are listed in Table 11, and applicable ASTM standards for mechanical testing

are listed in Table 12. More current information on mechanical testing of ceramics is provided in Ref 22.

Table 11 Typical mechanical properties of common ceramic materials

Young's modulus Flexural strength

Compressive strength

Material

GPa 10

psi MPa ksi MPa

ksi

Brick 5–20 0.7–2.9 5–10 0.7–1.5 10–25

1.5–3.6

Roof tile 5–20 0.7–2.9 8–15 1.2–2.2 10–25

1.5–3.6

Steatite 1–3 0.1–0.4 140–160 20–23 850–1000

123–145

Silica refractories, 96–97% SiO

… … 8–14 1.2–2.0 30–80

4.4–11.6

Fireclay refractories, 10–44% Al

20–45 2.9–6.5 5–15 0.7–2.2 10–80

1.5–11.6

Corundum refractories, 75–90% Al

30–120 4.4–17.4 10–150 1.5–22 40–200

5.8–30.7

Forsterite refractories 25–30 3.6–4.4 5–10 0.7–1.5 20–40

2.9–5.8

Magnesia refractories 30–35 4.4–5.1 8–200 1.2–29 40–100

5.8–14.5

Zircon refractories 35–40 5.1–5.8 80–200 12–29 30–60

4.4–8.7

Whiteware 10–20 1.5–2.9 20–25 2.9–3.6 30–40

4.4–5.8

Stoneware 30–70 4.4–10.2 20–40 2.9–5.8 40–100

5.8–14.5

Electrical porcelain 55–100 8.0–14.5 90–145 13–21 55–100

8.0–14.5

Capacitor ceramics … … 90–160 13–23 300–1000 44–145

Source: Ref 21

Table 12 ASTM standards related to mechanical testing of ceramics

Terminology

1145

Standard Definition of Terms Relating to Advanced Ceramics

1286

Standard System for Classification of Advanced Ceramics

Properties and performance (monolithic)

1161

Standard Test Method for Flexural Strength of Advanced Ceramics at Ambient Temperature

1211

Standard Test Method for Flexural Strength of Advanced Ceramics at Elevated Temperatures

1259

Standard Test Method for Dynamic Young's Modulus

1273

Standard Practice for Tensile Strength of Monolithic Advanced Ceramics at Ambient

Temperature

Design and evaluation

1175

Standard Guide to Test Methods for Nondestructive Testing of Advanced Ceramics

1198

Standard Test Method for Dynamic Young's Modulus

1212

Standard Practice for Fabricating Ceramic Reference Specimens Containing Seeded Voids

1239

Standard Practice for Reporting Uniaxial Strength Data and Estimating Weibull Distribution

Parameters for Advanced Ceramics

Characterization and processing

1251

Standard Guide for Determination of Specific Surface Area of Advanced Ceramics by Gas

Adsorption

1274

Standard Test Method for Advanced Ceramic Specific Surface Area by Physical Adsorption

1282

Standard Test Method for Determination of the Particle Size Distribution of Advanced Ceramics

by Centrifugal Photosedimentation

Ceramic composites

1275

Standard Practice for Monotonic Tensile Strength Testing of Continuous Fiber-Reinforced

Advanced Ceramics with Solid Rectangular Cross Section at Ambient Temperatures

Source: Ref 21

References cited in this section

21. Y.W. Mai and B. Cotterell, Int. J. Fract., Vol 32, 1986, p 105

22. C.R. Brinkman and G.D. Quinn, Standardization of Mechanical Properties Tests for Advanced

Ceramics, Mechanical Testing Methodology for Ceramic Design and Reliability, Marcel Dekker, 1998,

p 353–386

Mechanical Testing of Polymers and Ceramics

Room-Temperature Strength Tests

Uniaxial Tensile Strength. The nonductile nature of monolithic ceramics and their high sensitivity to stress

concentrators has meant that conventional direct tensile testing is difficult and expensive. Gripping with jaws,

screw threads, or other conventional devices causes invalid test results because of specimen breakage at the

grips. The high stiffness (elastic modulus) of many ceramics means that a misalignment of only a few

thousandths of a centimeter can lead to bending stresses with errors of 10% or more. Specimen preparation to

exacting tolerances with minimal machining damage and careful tapers to avoid stress concentrators has been

an expensive proposition. Considerable work has focused on improving tensile test methods for ceramics, with

the result that tensile testing is becoming more routine. Commercial equipment is readily available, and

specimen costs are falling. It will, however, always be more difficult to conduct direct tensile tests for ceramics

than for metals.

The experimental difficulties, coupled with the problems of fabricating sufficiently large specimens, have

prompted ceramists to use alternative test methods. The most common is flexure testing, in either the so-called

three-point or four-point configuration. The latter is usually further specified by a description of the distance

from the outer support points and the inner points, such as or four-point loading. The small size, low cost,

and easy preparation of a flexure specimen account for its popularity, but there are distinct drawbacks. The

bending creates a stress gradient in the specimen, and only a small volume is exposed to high tensile stress. The

specimens are very sensitive to edge or surface machining damage. The test appears easy to set up and conduct,

but misalignments and experimental errors can easily ruin it. Standard test methods are now available that

permit accurate strength measurements for standard sizes and shapes, as shown in Fig. 27.

Fig. 27 Flexure strength standard test methods; all dimensions in mm

Nevertheless, it is still preferable to perform direct tension testing. Current testing systems are designed with

self-aligning features that limit the imposed bending stresses to approximately 1%. There is usually less

extrapolation of the strength data from test specimen to component size. Tensile specimens are still expensive,

however, because of costly fabrication and machining. They are inconveniently large, as well, because most

systems are designed for high-temperature test rigs that use cold grips. Until recently, only a few laboratories

had the ability to test or to even afford direct tensile experiments. A new emphasis on attaining accurate, quality

data in support of ceramics in heat engine programs has led to rapid improvements in the field, and commercial

test systems are now readily available. Different tensile specimen geometries that are being used are shown in

Fig. 28 (Ref 23, 24, 25, 26, 27, 28).

Fig. 28 Tension specimens used for monolithic ceramics (each is in correct proportion to the others); all

dimensions in mm. Upper row for round specimens; lower row for flat specimens. Adapted from Ref 26

Another occasionally cited test for engineering ceramics is the so-called diametral compression test, or

Brazilian disk test, wherein a circular cylinder is loaded at its ends (Ref 29, 30). The test is actually biaxial,

because in addition to the tensile stresses that tend to laterally split the specimen, compressive stresses that are

three times as great act axially through the specimen. However, compressive stresses of this magnitude are not

likely to affect uniaxial strength, an effect peculiar to monolithic ceramics. The specimen loading is between

two platens with pads of compliant material (such as a metallic shim or paper) to avoid high shearing stresses.

Careful machining of the end faces of the specimen is essential, once again to avoid damage that compromises

the test. This point is often overlooked. This test is occasionally employed by ceramics processors for ceramics

fabricated in cylindrical shapes.

Many ceramic materials have strengths that are specific to the shaping process being used, such as injection-

molded turbocharger rotors or extruded heat-exchanger tubes. In such cases, it is not practical to cut tensile

specimens from the part, but separately cast tensile specimens may not have the same microstructure or defects

as the component and, therefore, are irrelevant (Ref 31, 32, 33). It is optimal to test components in as close a

configuration to the final component shape as possible. Thus, in the case of a tube, a ring can be cut from the

tube and pressurized to obtain a uniaxial hoop-stress-testing configuration (Ref 34, 35). Contrary to

expectations, such a test can be conducted at high temperatures. Indeed, one of the highest recorded strength

test temperatures for a ceramic (2180 °C, or 3955 °F) was on a pressurized tube (Ref 36). Extreme care must be

taken to ensure that the edges are not chipped and do not have excessive machining damage, lest the test merely

become a measure of machining damage.

There is no simple answer to the question of what specimen is best for measuring strength data. The best

practice is to test a configuration that most resembles the actual component in its service conditions and to

ensure that the test material accurately represents the component material. It is likely that the first available data

will be flexure-strength data, which are typically higher (10–50%) than tensile specimen data because of the

dependency of strength on test specimen size. Nevertheless, considering the tradeoffs in cost, quantity of

results, and difficulty in testing, it is likely that future engineering databases will feature complementary flexure

and tensile data. Indeed, it will be beneficial to have strength data from different sizes and shapes to permit an

assessment of material consistency, flaw uniformity, and the veracity of strength-size scaling models.

Elastic Modulus. Several methods are used to evaluate the elastic moduli of monolithic or fine-scaled, isotropic

composites. The most common are deflection measurements in flexural strength tests (with proper

consideration of the test machine compliance) or strain gage experiments in flexure or direct tension. Dynamic

measurements are also quite common, with either sonic excitation of prismatic specimens at their resonant

frequency or time-of-flight measurements of ultrasonic waves.

Interpretation of Uniaxial Strength. The scatter in uniaxial strengths is well modeled by Weibull statistics.

Weibull observed that the strength of brittle materials is controlled by the presence of randomly distributed

defects and and that failure is controlled by the largest, most severely stressed defect. Fracture occurs when a

defect in one particular element of the body reaches a critical loading. This analysis is colloquially known as

the weakest-link model, in direct analogy to the strength of a chain.

The Weibull modulus, m, has no units and is the factor that determines the scatter in strength. High values are

optimum. Traditional ceramics, such as whitewares and brick, may have values from 3 to 5. A good material

has a value that exceeds 10. A ceramic with an m value ≥30 has very consistent strengths and could be

practically considered to have a deterministic value of strength over a range of several orders of magnitude

volume.

Not only does strength scale with specimen size, but the magnitude of the change strongly depends on whether

the defects are surface or volume. Obviously, it is essential to know whether flaws are of one or the other

category if the laboratory strength data are going to be size-scaled to predict component performance.

A Weibull graph is a convenient means to report strength data. The graph usually has special axes chosen to

linearize the data. This is done in the same fashion that probability paper can be used to linearize data for a

Gaussian distribution.

The Weibull analysis is adequate for multiaxially, tensilely loaded ceramics, provided that the second or third

principal stresses are significantly less than the principal tensile stress. If this is not the case, then it is

appropriate to use more sophisticated analyses that take into account the effect of multiaxial tensile stresses on

defects. The Weibull analysis also has limitations if the defects are likely to grow subcritically during a test. A

newly recognized phenomenon that could occasionally pose problems in strength analysis is latent defect

caused by localized surface impact or contact stresses. Concentrated microdamage can occur that can lead to a

larger microcrack popping in after an incubation period (Ref 37, 38).

Strength values by themselves are only half the picture. The types of defects are equally important because each

flaw type has its own Weibull distribution, and because multiple flaw populations are common in ceramics.

Therefore, it is essential that the defects be as clearly associated with the strength values as possible.

Uniaxial Compression Strength. The high compressive strength of ceramics is a consequence of the resistance

of the material to plastic flow and the insensitivity of defects to compressive stress. Ancient structural

applications of ceramics were columns and walls that capitalized on high compressive strength. The fact that

ceramics fail at all in compression is a result of the distortion of the stress field in the immediate vicinity of the

tip of a defect. This distortion causes a localized tensile stress concentration that, for defects at the worst

orientation (-30° to the axial stress) is about ⅛ of the concentration if the specimen is loaded in tension. Thus, a

Griffith-type criterion for failure would predict that the compression specimen will fail at about 30° to the

specimen axial direction when the compressive stress is eight times the tension strength, but this is an over-

simplification.

The tensile stresses in the immediate vicinity of a defect will cause a crack to propagate stably for a slight

distance (Ref 39, 40). The crack then aligns itself with the compression stress and is arrested. Progressively

more defects grow until the damage that has accumulated in the specimen reaches some limit, and the specimen

virulently disintegrates into powder (often with a triboluminescent emission) (Ref 41, 42). Compression

strength thus depends not on the largest, worst-oriented, highest-stressed defect, but on the entire defect

population. The high compressive stresses can nucleate cracks due to twinning or dislocation activity, as well

(Ref 43, 44). Compression strength may have a dependence on the square root of grain size, because the size of

defects may scale with the average grain size or because of microplasticity in the grains. Weibull statistics are

irrelevant, and compression strengths often have very low scatter (Ref 41, 42).

The compression test appears deceptively easy to conduct. However, it is extremely difficult to accurately

measure compression strength, because slight misalignments can create bending stresses, and end loading

effects can cause parasitic tensile stresses that cause fracture. Mismatches of the elastic properties of the platens

and test specimen can cause tensile stresses or frictional constraints. Buckling can occur if specimens are too

long. Because the stresses being applied to a compression specimen are extremely high, the alignment errors

may be greater than they would be for an equivalent tensile test specimen. True compression tests may be as

difficult to conduct as direct tensile tests. Refined compression strength tests have been developed, as shown in

Fig. 29.

Fig. 29 Compression test specimens. P, applied load. Source: Ref 41 and 45

Fiber-Reinforced Ceramic Composites. Test methods for ceramic-matrix composites are typically quite

different than for monolithics, and they borrow heavily from organic-matrix and carbon-carbon test procedures.

Uniaxial tensile strength testing is most commonly done in direct tension or flexure loadings. Flexure testing is

not preferred, because the failure mechanism can be tensile, compressive, or interlaminar shear, depending on

the composite components, the reinforcement architecture, and the loading geometry. Flexure testing is

acceptable for measuring the matrix microcracking stress, the shear strength of one-dimensionally reinforced

composites (with the fibers perpendicular to the maximum stress), the effects of exposure or heat treatments, or

certain high-temperature properties.

Direct tensile loading is much easier to conduct for composites than for monolithics, because the former are

more tolerant of slight misalignments. Flat specimens with glued tabs (to avoid gripping damage) are quite

adequate with commercial test machine grips. Ordinary clip extensometers or strain gages are suitable for

measuring strain. At high temperatures, glued tabs are not adequate, and specimens with holes or tapered-

wedge shoulders are necessary. However, the low shear strength of unidirectionally reinforced composites can

make testing of pin-loaded specimens difficult or impossible, because the pins shear through the specimen. It is

actually easier to test two-dimensionally fiber-reinforced composites for this reason.

References cited in this section

23. D.F. Carroll, S.M. Wiederhorn, and D.E. Roberts, Technique for Tensile Creep Testing of Ceramics, J.

Am. Ceram. Soc., Vol 72 (No. 9), 1989, p 1610–1614

24. T. Soma, M. Matsui, and I. Oda, Tensile Strength of a Sintered Silicon Nitride, Nonoxide Technical and

Engineering Ceramics, Proc. of the International Conf., (Limerick, Ireland), S. Hampshire, Ed., 1985, p

361–374

25. T. Ohji, Towards Routine Tensile Testing, Int. J. High Technol. Ceram., Vol 4, 1988, p 211–225

26. G. Grathwohl, Current Testing Methods—A Critical Assessment, Int. J. High Technol. Ceram., Vol 4,

1988, p 123–142

27. K.C. Liu, H. Pih, and D.W. Voorhes, Uniaxial Tensile Strain Measurement for Ceramic Testing at

Elevated Temperatures: Requirements, Problems, and Solutions, Int. J. High Technol. Ceram., Vol 4,

1988, p 69–87

28. J. Nilsson and B. Mattsson, A New Tensile Test Method for Ceramic Materials, Ceramic Materials and

Components for Engines, W. Bunk and H. Hausner, Ed., German Ceramic Society, Berlin, 1986, p 651–

656

29. A. Rudnick, C.W. Marschall, W.H. Duckworth, and B.R. Emrich, “The Evaluation and Interpretation of

Mechanical Properties of Brittle Materials,” U.S. Air Force Technical Report TR 67-316, Air Force

Materials Laboratory, April 1968

30. J.E.O. Ovri and T.J. Davies, Diametral Compression of Silicon Nitride, Mater. Sci. Eng., Vol 96, 1987,

p 109–116

31. R. Morrell, Mechanical Properties of Engineering Ceramics: Test Bars versus Components, Mater. Sci.

Eng., Vol A109, 1989, p 131–137

32. G.D. Quinn and R. Morrell, Flexure Testing for Design of Engineering Ceramics: A Review, J. Am.

Ceram. Soc., Vol 74 (No. 9), 1991, p 2037–2066

33. S. Thrasher, Ceramic Applications in Turbine Engines (CATE) Program Summary, Proc. of the 21st

Contractors Coordination Meeting, Report P138, Society of Automotive Engineers, March 1984, p

255–267

34. R. Sedlacek and F.A. Halden, Method for Tensile Testing Brittle Materials, Rev. Sci. Instr., Vol 33 (No.

3), 1962, p 298–300

35. R. Jones and D.J. Rowcliffe, Tensile-Strength Distributions for Silicon Nitride and Silicon Carbide

Ceramics, J. Am. Ceram. Soc., Vol 58 (No. 9), 1979, p 836–844