ASM Metals HandBook Vol. 8 - Mechanical Testing and Evaluation

Подождите немного. Документ загружается.

silicate network becomes possible (Ref 114). With the network dissolution, atoms of modifiers such as

phosphorus and boron may enter the solution, producing a more acidic local environment.

Depending on the exchange rates between a restricted volume such as a crack and a bulk solution, local

environment composition inside this volume will be determined by an external environment (sufficient

exchange), glass composition (insufficient exchange), or both.

Crystalline Ceramics. It should be noted that environmental reactions for single crystals may depend on the

crystallographic orientation of an exposed surface. For the same environment/ceramic system, interaction may

be strongly dependent on the structure of a material. A situation is even more complicated for composites where

presence of second-phase particles may change susceptibility to an environmental attack.

Oxides. Environmental degradation of metal oxides such as alumina and titania has been observed in some

aqueous solutions (Ref 115).

Nonoxide Ceramics. Susceptibility to SCC for a variety of ceramics may be attributed to oxide removal by an

active environment. For example, silicon exhibits SCC when exposed to HF acid as revealed by indentation

induced cracking (Ref 116). For silicon-base ceramics, integrity and protective properties of silicon oxide may

also be critical in determining their environmental susceptibility. For example, SiC is sensitive to environmental

attack under a combination of elevated temperatures and vapors containing metal halides (Ref 117) or Na

in oxidizing atmospheres (Ref 118). With the Na

SO4 environment, corrosion was attributed to chemical

alteration of the normally protective SiO

. In aqueous acid solutions at ambient temperatures, susceptibility to

environmental attack may also depend critically on the formation of silica layers as demonstrated, for example,

by Cook et al. (Ref 119) in their study on sintered and reaction-bonded SiC. Here, both materials were passive

in aqueous solutions of HCl, HNO

, and H

, but reactive in HF solution. Here, reaction-bonded SiC

exhibited a higher degree of degradation. With the above, structural differences may be very important in

determining reactivity of a polycrystalline material. Grain-boundary chemistry would be very important notably

for nanocrystalline materials where the fraction of grain-boundary material increases substantially.

Environment/Fracture Zone Interactions

The crack-tip environment interactions may be divided into four categories that were originally suggested for

glasses (Ref 120), but may be extended to a broader range of ceramic materials.

Bond rupture reactions involve environmental assistance in breaking strained bonds at a near-tip region. With

all the complexity of mechanisms that may be involved in subcritical crack growth in ceramics (diffusion, Ref

121, surface relaxation, Ref 122, or plasticity, Ref 123), molecular reactions at the crack tip appear to be a

major driving force for subcritical crack growth. Michalske et al. (Ref 124) proposed a model of a reaction

between water and highly strained silica bonds in the near-tip region. Here, the ability of water to donate both

electrons and protons appears to be critical as shown in Fig. 26 (Ref 125). In excellent agreement with the

above model, similar crack velocity versus stress intensity (v/K) curves result from testing in water and other

species with the same structure and bonding characteristics as shown in Fig. 27 (Ref 125). Note that with the

nearly equal molecular sizes of water and ammonia, kinetic diagrams for these liquids are indistinguishable.

Fig. 26 Dissociative chemisorption reaction on crack-tip bond. (a) Physisorption on strained crack-tip

bond. (b) Concerted reaction resulting in crack-tip bond rupture. (c) Formation of silanol covered

fracture surface. Source: Ref 125

Fig. 27 Crack velocity versus stress-intensity diagram for vitreous silica at room temperature, showing

that water, hydrazine, ammonia, methanol, and formamide all have the effect of increasing the rate of

slow crack growth. Source: Ref 125

With crack growth rates dependent on the relative humidity rather than concentration of water in an

environment (Ref 126), it is implicit that there cannot be an absolutely inert liquid given even a small quantity

of water present. The above may be evident from comparison of kinetic curves for water and “dry” nitrogen as

shown in Fig. 27. Nearly equal initial slopes indicate similar mechanisms, while a difference between intercepts

comes from a difference in activities of water in different environments. In acidic aqueous solutions, the water

molecule is also suggested to be the key participant with the resulting mechanisms the same as in water.

Analogous mechanisms of water-assisted bond rupture may occur for crystalline materials such as alumina

having an ionic/covalent interatomic bond structure similar to silica (Ref 124). In basic solutions, the attack of

the OH ion appears to control crack growth rates.

Surface layer formation involves reactions changing the chemical composition and physical properties of a

ceramic surface. Surface layers with tensile (Ref 127) stresses may develop facilitating crack growth. Tensile

surface stresses may be responsible for a horizontal v/K dependence in the low K region for soda silica glass as

shown in Fig. 28. Note that absence of a static fatigue limit in this case may be quite disadvantageous for some

practical applications.

Fig. 28 Crack velocity versus stress-intensity diagram showing the stress-corrosion behavior of three

different silicate glasses in water. Source: Ref 127

Surface deposition involves preferential deposition of a reaction product in the crack-tip vicinity. For silica-

aqueous systems, dissolution-precipitation processes may lead to silica deposits at the crack tip resulting in

crack-tip blunting (Ref 127). On the other hand, in some cases, wedging associated with the reaction product

inside of a crack may add to the driving force for crack extension (Ref 128).

Intermolecular or Interatomic Surface Forces. Environmental effects may decrease interatomic forces or

produce surface bonds. Formation of hydrogen-bond linkages of surface adsorbed water molecules and

electrostatic bonds between anionic surface groups (Ref 127) may oppose crack propagation. In contrast, an

environment acting as a dielectric would decrease van der Waals dispersion forces (Ref 129). For highly ionic

solids such as MgF

, ion solvation, a mechanism involving neutralization of the electrostatic attraction between

ions may be a dominant mechanism (Ref 130) of static fatigue. Here, media with the high dielectric constants

such as water or methanol would be effective in facilitating subcritical crack growth.

Crack Growth Kinetics

Thresholds. In an ideal system with no additional shielding mechanisms, threshold corresponds to an

equilibrium where G = 2γ and frequencies of bond fracture and bond healing events are equal resulting in zero

crack velocity (Ref 131). Local processes in the crack-tip vicinity may affect local crack-tip stress, thus

changing local strain-energy release rates. For example, crack blunting due to selective silica deposition would

lower the local stress intensity. In contrast, tensile stresses in the near-surface region would increase the local

stress intensity. This mechanism results in the absence of a threshold for soda silica glass in water, as

mentioned earlier.

Reaction-Controlled Region. This part of the kinetic diagram is determined by environment, material

composition, and temperature.

Environment and Material Composition. In general, increasing pH results in decreasing the exponent N of Eq 1,

indicating greater susceptibility to SCC as shown in Fig. 29 for a pure silica glass (Ref 132). In this case, the

glass did not contain elements that might have affected local crack-tip environment via dissolution reactions.

With the more complex glass composition, the effect is more intricate since crack-tip environment could be

strongly affected by the glass composition. This effect may become prevalent at high crack velocities where

there is insufficient exchange with an external environment (Ref 133).

Fig. 29 Crack velocity as a function of pH for silica glass. Source: Ref 132

With moisture-containing “inert” media, the activity of water in a given medium will determine a horizontal

shift of the v/K curve with respect to that for water as shown in Fig. 27. As already mentioned in the previous

section and shown in Fig. 27, bonding and structure of the environmental molecule are very important in

determining SCC. Here, identical slopes of v/K curves for reactive media such as water, ammonia, hydrazine,

methanol, and formamide indicate similar mechanisms. It should be noted that intercepts of kinetic curves

increase with the increasing molecule sizes. The latter effect may possibly be related to crack-tip opening

limitations (Ref 125).

Temperature. With the thermally activated governing reaction, there is a strong temperature dependence on

crack growth rates as shown in Fig. 30 (Ref 131 and 134). Crack growth kinetics in the reaction controlled

region may be described with a recently developed reaction rate model (Ref 131).

v = (a dc/dA) v

sinh[(G - 2γ)/η]

(Eq 3)

Here, a is a characteristic length in the bond rupture process, G is strain-energy release rate, γ is surface energy

as affected by an environment, η = 2kT/a

, and v

is a parameter determined by material properties. This

parameter is crack-geometry independent and increases with increasing temperature. It is assumed that a crack

may be characterized by a single geometrical crack length parameter c, and this crack length parameter is

related to crack area by an expression: A α c

. Here, a parameter α = 1, 2, or 3 depends on the number of

dimensions for crack growth. For example, A = pc

for a circular crack geometry. As follows from Eq 3, a

parameter 2 γ determines the horizontal position of the ν/G curve, and 1/η characterizes the slope and νo, its

vertical position. For a given material/environment combination, increasing temperature would cause an

upward vertical displacement and decrease of the slope of the ν/G curve. These trends correspond to

experimentally observed temperature effects as shown in Fig. 30. A horizontal shift of the ν/G curves increasing

with increasing molecule size would also be predicted by Eq 3. Here, increasing molecule size would require

higher crack opening displacement and, consequently, an increasing G value.

Fig. 30 Crack growth rates for soda lime glass. Experimental data (Ref 134) are presented with the best

fit with the theoretical dependence given in Eq 3 (Ref 131)

Transport-Controlled Region. This region may depend on the activity of reactive species, for example, on a

relative humidity as shown in Fig. 31 (Ref 135). Crack geometry and pressure gradients along a crack may also

be important in determining transport towards a crack tip.

Fig. 31 Crack velocity curves for sapphire in moist air at different relative humidities. Double-cantilever

beam data (constant load) at 25 °C (77 °F). Solid curves are theoretical fits to experimental data. Source:

Ref 135

A transport model (Ref 131) describes crack velocity during stage II with the following expression:

(Eq 4)

where and and p

are the partial pressures of the reactive species in the external environment and at the

crack tip, respectively, m is the mass of the reactive species, and k is a dimensionless attenuation coefficient.

This parameter characterizes the constraining profile of the crack over which the pressure gradient acts.

Polycrystalline Materials. In addition to structure effects affecting ceramic/environment reactions, as discussed

earlier, presence of grain boundaries increase the level of complexity. Variation of grain-boundary chemistry

was correlated to the change of crack growth exponents for polycrystalline alumina as determined with

dynamic fatigue testing (Ref 136). Occurrence of transgranular versus intergranular fracture would depend on

variety of factors. These include crack velocity and chemical composition of grains and grain boundaries (Ref

137).

Testing Methods

Linear elastic fracture mechanics developed for metallic systems are generally applicable to ceramics, as

described in more detail in the article“Fracture Toughness of Ceramics and Ceramic Matrix Composites” in this

Volume. Classical fracture mechanics techniques for brittle materials are also discussed in Ref 138. Based on

the crack extension data, A and N can be determined by curve fitting Eq 1 to a power-law dependence. Among

disadvantages in the above method would be a large amount of material necessary to produce a specimen of

proper dimensions. A high scatter in experimental data, especially for polycrystalline ceramics, is a

disadvantage here.

Dynamic fatigue techniques involve measurement of strength as a function of stressing rate. Both of these

methods may be correlated with the following expression (Ref 139):

N + 1

= [2(N + 1) / AY

(N - 2)

(Eq 5)

Here A and N correspond to the same coefficient and exponent in Eq 1, Y is a constant, σ

is inert fracture

strength, is stressing rate, K

is critical stress-intensity factor. Alternatively, Eq 3 may be applied for the

analysis of dynamic fatigue results as described in (Ref 127). With this approach, parameters γ, η, and ν

determined from the best fit to experimental data may be used to predict crack growth kinetics for a specific

geometry of interest.

In general, there is a satisfactory correlation between crack growth parameters determined with classical

fracture mechanics and dynamic fatigue approaches. However, the role of surface state should be carefully

evaluated when interpreting experimental results obtained with the dynamic fatigue data (Ref 140). Also, for

reliable results, strength should be measured over at least three orders of magnitude in stressing rates (Ref 141).

Introducing a controlled-size flaw for dynamic fatigue testing would be advantageous especially with the flaw

size/geometry similar to those expected in service.

References cited in this section

110. S.M. Wiederhorn, S.W. Freimann, and E.R. Fuller, Jr., Effect of Water and Other Dielectrics on

Crack Growth, J. Mater. Sci., Vol 17, 1982, p 3460–3478

111. R.W. Douglas and T.M.M. El-Shamy, J. Am. Ceram. Soc., Vol 50, 1967

112. M.L. Hair, Hydroxyl Groups in Silica Surface, J. Non-Cryst. Solids, Vol 19, 1975, p 299–309

113. R.J. Charles, Glass Surface Analysis by Auger Electron Spectroscopy, J. Appl. Phys., Vol 2,

1958

114. C.G. Pantano, D.B. Dove, and G.Y. Onoda, J. Noncryst Solids, Vol 19, 1975, p 41–53

115. J.D. Venables, Review, Adhesion and Durability of Metal-Polymer Bonds, J. Mater. Sci., Vol

19, 1984, p 2431–2453

116. M.D. Thouless and R.F. Cook, Stress Corrosion Cracking in Silicon, Appl. Phys. Lett., Vol 56,

1990, p 1962–1964

117. J.I. Federer, Stress-Corrosion of SiC in an Oxidizing Atmosphere Containing NaCl, Adv. Ceram.

Mater., Vol 3, 1988, p 293–295

118. J.I. Federer, Corrosion of SiC Ceramic by Na

, Adv. Ceram. Mater., Vol 3, 1988, p 56–61

119. S.G. Cook, J.A. Little, and J E. King, Corrosion of Silicon Carbide Ceramics Using

Conventional and Electrochemical Methods, Br. Corros. J., Vol 29, 1994, p 183–189

120. B.C. Burkner and T.A. Michalske, Effect of Surface Corrosion on Glass Fracture, Fracture

Mechanics of Ceramics, Vol 8, R.C. Bradt, Ed., Plenum Press, New York, 1986, p 391

121. D.P.H. Hasselman, Ultrafine Grain Ceramics, J.J. Burke, N.L. Reed, and V. Weiss, Ed.,

Syracuse University Press, 1970, p 297–315

122. M. Tomozawa and Y.-L. Peng, Surface Relaxation as a Mechanism of Static Fatigue of Pristine

Silica Glass Fibers, J. Non-Cryst. Solids, Vol 240, 1998, p 104–109

123. D.M. Marsh, Proc. R. Soc. (London) A, Vol 270, 1964

124. T.A. Michalske and S.W. Freiman, A Molecular Mechanism for Stress Corrosion in Vitreous

Silica, J. Am. Ceram. Soc., Vol 66, 1983, p 284–288

125. T.A. Michalske and B.C. Bunker, Steric Effects in Stress Corrosion Fracture of Glass, J. Am.

Ceram. Soc., Vol 70, 1987, p 780–784

126. S.W. Freiman, Effects of Chemical Environments on Slow Crack Growth in Glasses and

Ceramics, J. Geophys. Res., Vol 89, 1984, p 4072–4076

127. B.C. Bunker and T.A. Michalske, Effect of Surface Corrosion on Glass Fracture, Fracture

Mechanics of Ceramics, Vol 8, R.C. Bradt, Ed., Plenum Press, 1986, p 391

128. C.J. Simmons and S.W. Freiman, Effect of Corrosion Processes on Subcritical Crack Growth in

Glass, J. Am. Ceram. Soc., Vol 64, 1981, p 683–686

129. D.R. Clarke, B.R. Lawn, and D.H. Roach, The Role of Surface Forces in Fracture, Fracture

Mechanics of Ceramics, Vol 8, R.C. Bradt, et al., Ed., Plenum Press, 1986

130. T.A. Michalske, B.C. Bunker, and S.J. Frieman, Stress Corrosion of Ionic and Mixed

Ionic/Covalent Solids, J. Am. Ceram. Soc., Vol 69, 1986, p 721–724

131. R.F. Cook, Environmentally Controlled Nonequilibrium Crack Propagation in Ceramics, Mater.

Sci. Eng. A, Vol 260, 1999, p 29–40

132. S.M. Wiederhorn and H. Johnson, Effect of pH on Crack Propagation in Glass, J. Am. Ceram.

Soc., Vol 56, 1973, p 192–197

133. S.M. Wiederhorn, Mechanisms of Subcritical Crack Growth in Glass, Fracture Mechanics of

Ceramics, p 549–581

134. S.M. Wiederhorn and I.H. Boltz, Stress Corrosion and Static Fatigue of Glass, J. Am. Ceram.

Soc., Vol 53, 1970, p 543–548

135. B.R. Lawn, Fracture of Brittle Solids, 2nd ed., Cambridge University Press, 1993, 378 pages

136. J.R. Hellmann, J. Matsko, S.W. Freiman, and T.L. Baker, Microstructure-Mechanical Property

Relationships in 94% Alumina Ceramics, Tailoring Multiphase and Composite Ceramics, R.E. Tressler,

G.L. Messing, C.G. Pantano, and R.E. Newnham, Ed., Plenum Press, 1986, p 367–380

137. J.J. Mechlosky and S.W. Freiman, Fractographic Analysis of Delayed Failure in Ceramics,

Fractography and Materials Science, STP 733, ASTM, 1981, p 246–258

138. M. Sakai and R.C. Bradt, Fracture Toughness Testing of Brittle Materials, Int. Mater. Rev., Vol

38, 1993, p 53–79

139. S.M. Wiederhorn, Fracture Mechanics of Ceramics, Vol 2, R.C. Bradt, D.P.H. Hasselmann, and

F.F. Lange, Ed., Plenum Press, 1974, p 613–646

140. J.E. Ritter, and R.P. LaPorte, Effect of Test Environment on Stress-Corrosion Susceptibility of

Glass, J. Am. Ceram. Soc., Vol 58, 1975, p 265–267

141. S.W. Freiman, Stress-Corrosion Cracking of Glasses and Ceramics, Stress-Corrosion Cracking,

Materials Performance and Evaluation, R.H. Jones, Ed., ASM International, 1992, p 337–344

Evaluation of Environmentally Assisted Crack Growth

Y. Katz, N. Tymiak, and W.W. Gerberich, University of Minnesota

Environmental Stress Cracking and Crazing in Polymers

Environmental stability of a polymer or an interface involving a polymeric material is of critical importance for

a variety of applications. Ranging from nanoscale films to pipelines, these include biomedical and

microelectronic devices, vessels, cladding, and composite materials. A variety of environments interacting with

polymers during processing/service conditions include organic and inorganic liquids, pressurized gases, living

tissues, and soils as well as irradiation exposure. In many cases, polymers are utilized to decrease the severity

of environmental attack on metallic components. For example, polymer films, coatings, and finishes are

designed to prevent environmental degradation of metallic surfaces. Polymer washers are frequently utilized in

pipelines in order to eliminate galvanic contacts. Similarly, polymer fillers serve to reduce crevice corrosion

effects at joint sites. For any of the above, the preventive effect is determined by the structural integrity of a

polymeric component under exposure to an aggressive environment (with or without mechanical stresses).

Design considerations for a polymer as a prospective material for a given application should account for:

• How interaction with a given environment affects the polymer microstructure on the molecular scale

• How stress-strain conditions particular to a given application influence these interactions

• How resulting microstructural changes affect plasticity and fracture behavior of a polymer

Note that consideration should be given to the highly time-dependent and temperature-dependent mechanical

behavior of polymers as viscoelastic-plastic solids. This introduces an additional challenge if combined with the

time-dependent nature of deformation/environment interactions.

Deformation and Fracture with No Environment

Factors determining mechanical behavior of polymers include structure, applied stress-state, testing time, and

temperature with respect to T

, the glass-transition temperature. With an increasing degree of crystallinity and a

temperature sufficiently below a glass transition, polymers experience a ductile-brittle transition. Plastic

deformation in polymers occurs either via shear yielding, crazing, or plastic instability (necking). Conditions

essential for craze formation are temperature above T

and a dilatational hydrostatic stress-state component.

Dilatational stress results in formation of highly voided regions with so-called fibrils. These are fiberlike

semicohesive structures capable of sustaining significant stress levels. In many cases, craze zone evolution at

the vicinity of a nucleating or advancing crack plays an essential role in the fracture process in amorphous (Ref

142) and crystalline (Ref 143) polymers.

Polymer/Environment Interactions with No Stress

The main issues dealing with solid polymer/environment interactions are the penetration mechanism (diffusion)

and change in polymer properties determining its mechanical behavior. Possible environmentally related

changes include change in degree of crystallinity, polymer surface energy reduction (Ref 144), decrease of T

(plasticizing), and dimensional change of a polymer (swelling). The latter change can be expressed in terms of

the coefficient of hydroelasticity. This parameter is sensitive to both the chemical composition and the physical

structure of a polymer (Ref 145). Finally, dissolution of a polymer may occur.

Environmentally Assisted Cracking and Crazing

Several important points need to be emphasized here, as shown in Fig. 32. First of all, environmentally

enhanced cracking and/or crazing may occur even in “inert” media that does not affect a given polymer under

no applied stress (Ref 146). Moreover, with the absence of environmental effects under simple tension,

environmental stress crazing and/or cracking may occur under more complex stress states (Ref 147). For active

media, stress may enhance environmental effects such as plasticization and swelling. This may become

critically important at the highly stressed crack-tip region. In addition, for fast-moving cracks (200–650 m/s, or

650–2130 ft/s), a crack-tip temperature increase up to 500 K is possible (Ref 148). Thus, even for medium

growth rates with a small environmentally induced decrease in T

and/or a small temperature increase, local

effects at the crack tip may be significant. Organic plasticizers or even inorganic salt solutions (Ref 149) are

known to affect tensile properties of some polymers, for example, nylon. Local stress-enhanced crack-tip

environmental interactions may induce a substantial decrease in the Young's modulus, thus easing local

deformation. With the above examples, it becomes evident that predictions for environmental crazing or

cracking for a given polymer/environment system may not be based exclusively on the interactions with no

stress applied. Also, local crack-tip conditions need to be accounted for. It should also be noted that both

cracking and crazing kinetics must be considered in assessing environmental effects. For example, high craze

rates but more stable craze structure may result in longer time to complete failure. A brief consideration of the

mechanisms of environmental stress crazing and cracking is given here as well as references for most

commonly used experimental procedures. Examples for selected polymer/environment systems are included

with consideration for dominating mechanisms and critical parameters.

Fig. 32 Polymer/environment/stress interactions

Mechanisms in Liquid Media

Liquid Transport through a Porous Craze Zone or a Crack. Suggested driving forces are: (a) atmospheric

pressure as the craze voids are under a vacuum, (b) capillary pressure in the craze or a crack, and (c)

superimposed hydrostatic pressure (Ref 150). Viscosity of the liquid and a spreading coefficient with respect to

a polymer are critical parameters controlling liquid transport through a crack and/or a craze zone. Specific

liquid/polymer interactions may also be involved here.

Liquid Fracture Zone Interactions. An active media absorption at the fracture zone surface is followed by local

changes in polymer properties. For some polymers, for example, polyethylene, stress enhances liquid

absorption. Thus, under exposure to a solvent, enhanced swelling is expected in the highly stressed near-tip

regions. It seems that solvent swells and plasticizes both the craze fibrils and the layer of bulk polymer

contiguous to the craze surface. One of the consequences is believed to be a stress relaxation in a fibril. As the

stress relaxes in a fibril, new fibrils need to be generated to balance the stress concentration at the crack tip.

Thus, the craze tip advances. There is some evidence that inorganic media such as metal salt solutions can also

diffuse into polymer and cause stress relaxation in a fibril (Ref 149). It is suggested that for media resistant to

swelling, local plasticization may increase the effective radius of curvature at the crack tip, thus reducing the

stress concentration (Ref 151).

Craze and/or Crack Growth Kinetics. For a variety of glassy polymer/liquid systems, crack growth rate versus

stress-intensity dependencies exhibit three characteristic regions (Ref 147). In region I, crack growth rate

increases rapidly with the stress-intensity factor once a threshold value is exceeded. This region is believed to

be controlled by environment/polymer interactions. In region II, crack growth rate is controlled either by

hydrodynamic flow properties of the environment (Ref 152), or by velocity of adsorption at the fracture process

zone (Ref 153). For some systems, a region III exists where the fracture process is controlled by the mechanical

properties of a material in air (Ref 154). However, in some cases, K

in the presence of an environment may be

either decreased or increased. For example, environmentally enhanced crazing results in an increased K

for

polymethylmethacrylate (PMMA) exposed to organic solvents (Ref 147). In this case, K

may also depend on

crosshead speed since the amount of crazing formed at the tip can vary with crack velocity (Ref 155). At the

same time, the reader is cautioned to be aware that such solvents could enhance crack growth kinetics well

below K

. If a solvent causes dissolution and not crazing, K

can be reduced (Ref 147). In some cases, crack

deceleration due to crack-tip blunting is observed following region II (Ref 156). It should become clear that

different solvent/polymer interactions may produce enhanced or decreased crack growth rates or final fracture,

requiring case-by-case evaluation. A level of complexity increases substantially if a polymer is subjected to a

cyclic rather than monotonic load. Here, crack growth may be highly dependent on the test frequency and

loading amplitude as well as the stress-intensity factor range (Ref 157).

Mechanisms in Gaseous Media

Gas Fracture Zone Interactions. In general, hydrostatic pressure will increase T

. However, if the pressurizing

environment is soluble in the polymer, one might expect an initial decrease in T

as a polymer is plasticized by

an environment (Ref 158). An effect of gases such as argon, oxygen, and carbon dioxide close to their

condensation point (high thermodynamic activity) is to reduce the surface energy of a crystalline polymer,

facilitating the creation of a new surface in the holes and voids of the craze (Ref 159). On the other hand, high

dilatational stress at the tip of an incidental flaw or an existing craze promotes enhanced absorption of a gas.

Craze Growth. The locally absorbed gas acts as a plasticizer, easing the flow involved in the craze nucleation

and growth. The above effects are suggested to explain the ductile behavior of crystalline polymers between

liquid nitrogen and room temperature. In the absence of a gas or at temperatures sufficiently above a gas

condensation point, the crystalline material is rather brittle with a very limited extent of crazing. Craze velocity

is directly related to the diffusion coefficient of a gas in a polymer (Ref 160).

Testing

Smooth specimens are commonly used when craze initiation and growth are investigated (Ref 150). If

diffusivity of an environment is high, a constant strain rate tensile test may be utilized (Ref 150, 156). For low-

diffusivity liquids, a creep test may be a better option (Ref 156). Note that for some inert liquids, no crazing