8/12/2019 Stress Corrosion and Static Fatigue of Glass http://slidepdf.com/reader/full/stress-corrosion-and-static-fatigue-of-glass 1/6 October 1970 Stress Corrosion and Static Fatigue of Glass 543 to be published in Zeitschrift fuer Anorganische und Allge- meine Chemie. b) Vladimir Balek, “Rare-Gas Diffusion, Reactivity and Elec trical Conductivity of Solid ZnO”; to be published in Phys- ica Status Solidi. l2 J Beretka and M. J. Ridge, “Formation of Zinc Ferrite at Low Temperatures,” Nature (London), 216 [5114] 473-74 1967). l Vladimir Balek. “Emanation Method for Estimating Re- activity of Ferric Oxides Prepared from Different Sou;ces,” J. AppZ. Chenz., 20 [3] 73 1970). i4 Eestmir Jech and Roger Kelly, “Bombardment-Induced Disorder: (a) S Malcic, Lj. Petrovic, and S, J. Kiss, “X-Ray In- vestigation of Nickel-Zinc Fe rrite Formation,” Bull. Boris Kid- ric Inst. Nucl. Sci., Ceram. Met., 20 [l] 1969). I,” 3 Phys. Chem. Solids, 30 [3] 465-74 1969). vv V b) D. Cerovic, I Mom&ovi&, and S J. Kiss; pp. 443-46 in Proceedings of the Vth International Congress on X-Ray Optics and Microanalysis, Tubingen 1968. Edited by G. MOI- lenstedt and K. H. Gaukler. Spring er-Verl ag, Berlin, 1969. lap Reijnen; pp. 562-71 in Reactivity of Solids. Edi ted by G. M. Schwarb. American Elsevier Publishing Co., Inc., New York, 1965. Vladimir Balek; Ph.D. T hesis, Moscow Stat e University, 1967. Vladimir Balek, “Temperature Dependence of Characteris- tic Prope rties of ci-Fez03 Powders”; to be published in Journal of Materials Science. I’ a) Roland Lindner. “Diffusion of Radioactive Zinc in ZinGIfon Spinel and Zinc Oxide,” Acta Chem. Scand., 6 [4] (b) Roland Lindner, “Formation of Spinels and Silicates by Reactions in the Solid State, Investigated by Method of Radio- active Tracers,” Z. Elektrochem., 59 [lo] 967-70 1955). 457-67 1952). Stress Corrosion and Static Fatigue of Glass S. M. WIEDERHORN and L. H. BOLZ Institute for Materials Research, National Bureau of S tandards, Washington, D. C. 20234 Stress corrosion cracking of six glasses was studied using fracture mechanics techniques. Crack velocities in water were measured as a function of applied stress intensity factor and temperature, and apparent activation energies for crack mo- tion were obtained. Data were consistent with the universal fatigue cu rve for static fatigue of glass, which depended on glass composition. Of the glasses tested, silica glass was most resistant to static fatigue, followed by the low-alkali alurnino- silicate and borosilicate glasses. Sodium was detrimental to st re ss corrosion resistance. The crack velocity data could be explained by the Charles and Hillig theory of stress corrosion. It is probable that stress corrosion of glass is normally caused and controlled by a chemical reaction between the glass and water. I. Introduction LASS is noted for its chemical inertness and general re- G sistance to corrosion; therefore, it is used in the chemical industry and in the laboratory when chemical inertness is required. Despite this well-known property, glass is extremely susceptible to stress corrosion cracking caused by water in the environment.’-’ This phenomenon is known in the glass litera- ture as static fatigue or delayed failure. The susceptibility of glass to stress corrosion cracking was observed first by Grenet,4who noted a time delay to failure and a loading ra te dependence of strength. Although he was unable to explain his observations, subsequent studies have demon- strated that the effect is an activated process caused by water in the environment?-’ It is currently believed that static fatigue of glass res ults from the growth of small crac ks in the surface of glass under the combined influence of water vapor and applied A new methodi0-“ for studying the stress corrosion of glass involves measuring the velocity of macroscopic cracks as a Presented at the 71st Annual Meeting, The American Ce- ramic Society, Washington, D. C., May 7, 1969 (Glass Divi- sion, No. 474-69). Received March 2, 1970; revised copy received April 16,1970. Supported by the Advanced Research Projects Agency of the Department of Defense. Fig. 1. Specimen configuration. Cross-hatched area desig- nates fracture surface; direction of propagation is from right to left. Web is not shown. function of external and internal variables such as temper- ature, applied load, and composition. Basic parameters such as activation energies and volumes may be obtained by this technique. These parameter s in turn can be related t other rate processes that may occur at the crack tip during growth and may in fact control the rate of stress corrosion. Thus, information obtained from crack growth studies may be used to describe stress corrosion in terms of fundamental processes such as diffusion and chemical reactions. The present work is a study of crack growth rates in six glasses tested in water with temperature and applied load as external variables. The data are consistent with the stress corrosion theory of Charles and Hillig17~1s nd with static fatigue data in the form of the universal fatigue curve obtained by Mould and Southwick.“ II. Experimental Procedure The experimental geometry is shown in Fig. 1. The essen- tial experimental parameters are crack length, L; hickness, w; height, t; and applied load, P. Crack propagation was re- strained to the midplane by using slotted cantilever specimens 75 by 25 by 1.5 mm in which two slots 0.3 mm deep and 0.1 mm wide were cut along the midplane of the specimen, leaving
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
8/12/2019 Stress Corrosion and Static Fatigue of Glass
October 1970 Stress Corrosion and Static Fatigue of Glass 543
to be published in Zeitschrift fuer Anorganische und Allge-meine Chemie.
b) Vladimir Balek, “Rare-Gas Diffusion, Reactivity andElectrical Conductivity of Solid ZnO”; to be published in Phys-ica Status Solidi.
l2 J Beretka and M. J. Ridge, “Formation of Zinc Ferriteat Low Temperatures,” Nature (London), 216 [5114] 473-741967).l Vladimir Balek. “Emanation Method for Estimating Re-
activity of Ferr ic Oxides Prepared from Different Sou;ces,”
J . AppZ. Chenz., 20 [3] 73 1970).i4 Eestmir Jech and Roger Kelly, “Bombardment-Induced
Disorder:
(a) S Malcic, Lj. Petrovic, and S, J. Kiss, “X-Ray In-vestigation of Nickel-Zinc Ferrite Formation,” Bull. Boris Kid-ric Inst. Nucl. Sci., Ceram. Met., 20 [l] 1969).
I,” 3 Phys. Chem. Solids, 30 [3] 465-74 1969).v v V
b) D. Cerovic, I Mom&ovi&, and S J. Kiss; pp. 443-46in Proceedings of the Vth International Congress on X-RayOptics and Microanalysis, Tubingen 1968. Edited by G. MOI-lenstedt and K. H. Gaukler. Springer-Verlag, Berlin, 1969.lap Reijnen; pp. 562-71 in Reactivity of Solids. Edited by
G. M. Schwarb. American Elsevier Publishing Co., Inc., NewYork, 1965.
Vladimir Balek; Ph.D. Thesis, Moscow State University,1967.
Vladimir Balek, “Temperature Dependence of Characteris-tic Properties of ci-Fez03Powders”; to be published in Journalof Materials Science.
I’ a) Roland Lindner. “Diffusion of Radioactive Zinc inZinGIfon Spinel and Zinc Oxide,” Acta Chem. Scand., 6 [4]
(b) Roland Lindner, “Formation of Spinels and Silicates byReactions in the Solid State, Investigated by Method of Radio-active Tracers,” Z. Elektrochem., 59 [lo] 967-70 1955).
457-67 1952).
Stress Corrosion a n d Static Fatigue of GlassS. M. WIEDERHORN and L. H. BOLZ
Institute for Materials Research, National Bureau of Standards, Washington, D. C. 20234
Stress corrosion cracking of six glasses was studied using
fracture mechanics techniques. Crack velocities in water were
measured as a function of applied stress intensity factor and
temperature, and apparent activation energies for crack mo-
tion were obtained. Data were consistent with the universal
fatigue cu rve for stat ic fatigue of glass, which depended on
glass composition. Of the glasses tested, silica glass was most
resistant to static fatigue, followed by the low-alkali alurnino-
silicate and borosilicate glasses. Sodium was detrimental tostress corrosion resistance. The crack velocity data could be
explained by the Charles and Hillig theory of stress corrosion.
It is probable that stress corrosion of glass is normally caused
and controlled by a chemical reaction between the glass and
water.
I. Introduction
LASS is noted for its chemical inertness and general re-G sistance to corrosion; therefore, it is used in the chemical
industry and in the laboratory when chemical inertness is
required. Despite this well-known property, glass is extremely
susceptible to stress corrosion cracking caused by water in the
environment.’-’ This phenomenon is known in the glass litera-
ture as static fatigue or delayed failure.
The susceptibility of glass to stress corrosion cracking was
observed first by Grenet,4who noted a time delay to failure and
a loading ra te dependence of strength. Although he was unable
to explain his observations, subsequent studies have demon-
strated that the effect is an activated process caused by water
in the environment?-’ It is currently believed that static
fatigue of glass results from the growth of small cracks in the
surface of glass under the combined influence of water vapor
and applied
A new methodi0-“ for studying the stress corrosion of glass
involves measuring the velocity of macroscopic cracks as a
Presented at the 71st Annual Meeting, The American Ce-ramic Society, Washington, D. C., May 7, 1969 (Glass Divi-sion, No. 474-69). Received March 2, 1970; revised copyreceived April 16, 1970.
Supported by the Advanced Research Projects Agency ofthe Department of Defense.
Fig. 1. Specimen configuration. Cross-hatched area desig-nates fracture surface; direction of propagation is from
right to left. Web is not shown.
function of external and internal variables such as temper-
ature, applied load, and composition. Basic parameters such
as activation energies and volumes may be obtained by this
technique. These parameters in turn can be related t other
rate processes that may occur at the crack tip during growth
and may in fact control the rate of stress corrosion. Thus,information obtained from crack growth studies may be used
to describe stress corrosion in terms of fundamental processes
such as diffusion and chemical reactions.
The present work is a study of crack growth rates in six
glasses tested in water with temperature and applied load as
external variables. The data are consistent with the stress
corrosion theory of Charles and Hillig17~1snd with static fatigue
data in the form of the universal fatigue curve obtained by
Mould and Southwick.“
II. Experimental Procedure
The experimental geometry is shown in Fig. 1. The essen-
tial experimental parameters a re crack length, L; hickness, w ;height, t; and applied load, P. Crack propagation was re-
strained to the midplane by using slotted cantilever specimens
75 by 25 by 1.5 mm in which two slots 0.3 mm deep and 0.1
mm wide were cut along the midplane of the specimen, leaving
8/12/2019 Stress Corrosion and Static Fatigue of Glass
October 1970 Stress Corrosion and Static Fatigue of Glass 547
I I 1 I
2
.n
0X
E
Y
\
z
InIn
LT
in
wv
I - S I L I C A
2-ALUMINOSIL ICATE
3 - B O R O S IL ICAT E
4 AL U M IN O SIL IC AT E
5 - S O D A - L I M E
6 L E A D - A L K A L I
tI
I0 1 0 2 Io3 104
F A IL UR E T IM E ,s e c o n d s
Fig. 6. Static fatigue curves for glass compositions studied.Curves calculated from Eq. 4) using data in Table 11.
Figure 6 shows that fused silica has the greatest stress
corrosion resistance for long time loads, whereas the soda-
lime silicate and lead-alkali glasses have the poorest resistance.
The aluminosilicate I and borosilicate glasses occupy inter-mediate positions, with aluminosilicate I appearing superior.
The behavior seems to be related to the sodium content of the
glasses; those which contain large amounts of sodium behave
poorly under stress corrosion conditions. As the temperature
is increased, the resistance of all the glasses to stress corro-
sion decreases. The relative positions of the curves of Fig. 6
remain approximately the same, and conclusions concerning
the relative merits of the glasses for stress corrosion resistance
are unchanged.
3 ) Characterist ic Durat ions
calculated from Eq. 4)
and those obtained experimentally by Mould and So~thwick'~
will now be compared. The characteristic duration is given
by Eq. 4)when Ki=Krc/2 and v I = v u 5 = v o xp ( - E + b K d2 ) / R T . Except for L, all terms on the right side of Eq. 4)are constants at a given temperature. For a Griffith crack,
K l c . = a N V Z , and for a penny-shaped crack, KrC=20N x
Substituting these equations into Eq. 4) ives
The characteristic durations, to
to s = 4/~)RTKrc/bvu.a)~N-~ 6 )
7 )
for the penny-shaped crack. The characteristic duration is
proportional to the inverse square of the strength in liquid
nitrogen, and, for a given strength, the duration of the penny-
shaped cracks is greater than that of the Griffith cracks.
The characteristic duration data of Mould and Southwick are
presented in Fig. 7. The data are plotted as log to vs log uN
for the Griffith-type crack and
to ,=T RTKrc/bVu 0 UN-*
Originally, Mould and Southwick plotted their data as logt o . vs ( 1/U )2.
I I I I I I
T
S LO P E =- 6 . 5
I I I I I I
I O Q I O ~ N
4.0 41 4.2 4.3 4.4 4.5
Fig. 7.subjected to mechanical damage treatments.
Characteristic duration of soda-lime silicate glassData taken
from Ref. 19.
for easy comparison with Eqs. 6) and 7).* As was dis-
cussed by Mould and Southwick, different curves ar e obtained
for slides containing point flaws and linear flaws. The slopes
of both curves a re approximately -6.5; the curve for pointflaws lies at longer characteristic durations than the other. If
the penny- and Griffith-shaped cracks are assumed to be repre-
sentative of point and linear flaws, respectively, the displace-
ment of the calculated curves agrees with that found by Mould
and Southwick. The slopes of the two sets of data, however, do
not agree, since the slope of the experimental data (Fig. 7 ) is
about 3 times that determined from Eqs. 6) and (7). Thus,
the measured and calculated durations do not agree quantita-
tively. Although Eq. 4)may be used to judge the relative
resistance of glass to static fatigue qualitatively, it does not
predict the failure times quantitatively.
The difference between the calculated and measured char-
acteristic durations may result from differences in the type of
cracks studied. Cracks introduced by the techniques used by
Mould and Southwick were probably representative of flaws
found in real materials and consequently were much more ir-
regular than those studied in the present investigation. It is
doubtful that cracks introduced by abrasion or impact are flat
or lie perpendicular to the tensile axis. Therefore, the de-
pendence of stress intensity factor on flaw size may differ from
those used in the present study. Considering these factors, the
measured and calculated characteristic durations might be
expected to disagree. The universal fatigue curves seem
much less sensitive to these effects, as evidenced by the good
agreement obtained in Fig. 5.
4 ) Stat ic Fat igue as a Chemical Process
The crack velocity data presented in the present paper agree
very well with the stress corrosion theory developed by Charles
and Hillig,i'*'8 who assumed that stat ic fatigue in glass i s con-
trolled by a chemical reaction between the glass and water in
the environment. Since chemical reactions ar e activated proc-
esses, static fatigue should also be an activated process. Also,
8/12/2019 Stress Corrosion and Static Fatigue of Glass
548 Journal of The American Ceramic Society-Wiederhorn and Bolz Vol. 53, No. 10
the act ivat ion energy for the process will be stress-sensitive,and the react ion would be expected to occur most rapidly wherethe s tress f ields were the greatest . Thus, the water is expectedto react most rapidly at the crack t ip , extending the cracklength until the Griffith conditions for spontaneous failure aresatisfied.
Charles and Hillig developed these ideas into a quantitativetheory fo r stat ic fatigue, in which the c rac k velocity is given by
v =vo exp ( -Ef+V b-Vnfy/p) /RT 8 )
where Et is the s tres s f ree act ivat ion energy, Vr the activationvolume, u the s t r ess a t the crack t ip , VM the molar volume of
the glass , y the interfacial surface energy between the glassand the react ion products , and p the radius of curvature of thecrac k t ip .
Equation (8) was der ived for a two-dimensional Griffith
crack for which 0=2RI/V= At stresses greater than thefat igue l imit , crack sharpening occurs, and p decreases to a
small value l imited by th e s tructure of the glass. Thus, abovethe stat ic fat igue l imit , the third term is a constant independentof applied load, an d Eq . (8) t akes the fo rm
v = v u e x p ( -E*+2VtKI/V/?rp)/RT (9)
where E” = E t f V M y / p . Depending on the values assumed fory an d p , VMy/p r an g es f r o m 10 to 20% E“. Equation 7) i s
identical to Eq. (2) if Vf - ( b / 2 ) V g Thus, the observedform of the exper imental data for large loads is identical tothat predicted f rom the theory of Charles and Hillig. The bend-ing of the cu rves fo r soda-lime silicate an d borosilicate glas sesat low loads may indicate a change in crack t ip radius, a
change in rate- l imit ing mechanism, or an approach to thefatigue limit.
V. SummaryThe stress corrosion of glass was studied using f racture
mech anics techniques. Crac k velocity da ta were obtained as
a function of glass com position, applied stre ss intensity facto r,and temperature. The observed crack velocity data dependedstrongly on glass composition. The stress corrosion behaviorwas consistent with s tat ic fat igue studies of glass since uni-versal fat igue curves determined f rom the crack velocity dataagreed with those obtained f rom strength data. Universalfatigue curves depended on glass composition, confirming theobservations of other workers. Calculated static fatigue curve scould be used to judge which glasses would exhibit good stress
corrosion resistan ce and which would not. Of the glassesstudied, silica glass had the greatest stress corrosion resist-ance, followed by the low-alkali alumino silicate and borosilicateglasses. Sodium seemed to be detrimental to the s tress cor-rosion resistance of glass. Calcula ted cha racteris tic dura-
t ions agre ed only quali tat ively with those measured by M ouldand Southwick for soda-lime silicate glass. The difference inbehavior wa s at tr ibuted to dif ferences in crack sha pe in the
exper iments . The cra ck velocity dat a presented in the pres-ent paper agreed with the stress corrosion theory of Charlesand Hillig; it is probable that the stress corrosion of glass iscaused and controlled by a chemical react ion between waterin the environment and the glass .
References
‘ R . J Charles; pp. 1-38 in Progress in Ceramic Science,Vol. 1. Edi ted by J E. Burke. Pergam on Pres s, New York,1961.
a W. B. Hillig; pp. 152-94 in Modern Aspects of the VitreousState, Vol. 2. Edited by J. D. Mackenzie. Butt erw orth Inc.,Washineton. D. C.. 1962.
C. JrPhii lips, “Strength a nd Weakness of Brittle Materials,”
Louis Grenet. “Mech anical Stre ngth of Glass.’’ Bull. SOC.Ay er . Sci. , 53 [l] 20-51 1965).
Etc. Industr. Nut . Par is (Ser . 5), 4,33 8-48 1899j.
S y . G lass Technol. , 13 [52] 351-60T 1929).L. H. Milligan, “Strength of Glass Containing Cracks,” J .
C. Gurnev and S . Pearson. “Effect of Surroundine Atmos-phe re on Defayed Fr ac tu re of Glass,” Pro c. Phys . Soc.,London,
Sect. B , 62 [356] 469-76 1949).‘T. C. Baker and F. W. Preston. ‘‘Fatigue of Glass UnderStat ic Loads,” J . Appl. Phys., 17 [3]’170-78 71946) .
T. C . Baker and F. W. Preston, “Effect of Wa ter on Strengthof Glass,” ibid., pp. 179-88.’ . J. Charles , “Stat ic Fatigue of Glass: I, 11 ” ibid., 29
[111 1549-60 1958).G. R. Irwin, “Moisture Assisted Slow Crac k Extension in
Glass Plates ,” NRL Memorandum Repor t 1678, J an u a r y 28,1966.‘IS. M. Wiederhorn; pp. 503-28 in Materials Science Re-
search, Vol. 3. Edited by W. W. Kriegel and H ayne Palm ourIII , Plenum Press, New York, 1966.
S. M. Wiederhorn; pp. 293317 in Environment-SensitiveMechanical Behavior of Materials. Edite d by A. R. C. West-wood and N. s. Stoloff. Gordon an d Bre ach, S cience Publishers,In$, New York, 1966.
S. M. Wiederhorn. “Influence of Water Vavor on CrackPropagation in Soda-Lime Glass,’’ J . Amer. C6ram. SOC.,0
l4 S M. Wiederhorn. “Moisture Assisted Crack Growth in[8] 407-14 196 7).
Ceramics,” Int. J . Fib Mech., 4 [2] 171-77 1968).16K. Schonert, H. Um hauer, a nd W. Klemm , “Influence of
Temperature and Environment on Slow Crack Propagation inGlass”; Paper 41, Prepr int of Second International ConferenceoygF ractu re, Brighton, April 13-18, 1969.
J. A. Kies and A. B. J Clark, “Fracture PropagationRa tes a nd Times to Fa il Following Proof Stress in Bulk Glass”;P:per 42 in Ref. 15.
R. J. Ch arles an d W. B. Hillig; pp. 511-27 in Symposiumon M echanical Strength of Glass an d Ways of Improving It.Florence, I taly , September 25-29, 1961. Union ScientifiqueC2:tinentale du Ve rre, Cha rleroi, Belg ium , 1962.
W. B. Hillig and R. J Charles; pp. 682-705 in High-StrengthMater ials . Edite d by V. F Zackey. John Wiley Sons, Inc.,
New York, 1965.lD R. E. Mould and R. D. Southwick, “Strength and Static
Fatigue of Abraded Glass Under Controlled Ambient Condi-tions: 11.” J . Amer. Ceram. SOC.. 2 r121 582-92 1959\.’’K. R.’ Linger and D. G. Holloway,L‘‘I?racture En e& y of
’ J. E. Srawlev and Bernard Gross. “Stress Intensitv FactorsGlass,” Phil. Mag., 18 [156] 1269-80 1968).
for Crackline-Loaded Edge-Crack Specimens,” MGer. Res.Stand., 7 [4] 155-62 1967).
S . M. Wiederhorn, A. M. Shorb, and R. L. Moses, “CriticalAnalysis of the Th eory of t he D ouble Cantilever Method ofMeasur ing Fracture-Surface Energies,” J . Appl. Phys., 39 [3]
23 P. C. Paris and G. C. Sih. “Stress Analvsis of Cracks”:1569-72 1968).
p p 30-81 in Fracture Toughness Test ing and i t s Applications,AS rM Special Techn ical Publication No. 381. American Societyfoz Testing and Mater ials , Philadelphia, Pa. , 1965.
A. S. Tetelman and A. J. McEvily, Jr . , Fracture of Struc-tw a l Mate rials. John Wiley Sons, Inc., New York, 1967.
S. M. Wiederhorn; pp. 217-41 in Mechanical and ThermalProp erties of Ceram ics. Edite d by J. B. Wachtman, Jr. Nut.B y . S ta nd . US) pec. Publ., 1969, No. 303; 268 pp.
A. J. Holland and W. E. S. Turner, “Effect of SustainedLoading on Breaking Strength of Sheet Glass,” J. SOC.GlassT z h o l . ,24 [ lo ,] 46-57T 1940).
E. B. Shand, “Expe rimental Study of Fr ac tu re of Glass:I,y8J.Amer. Cera,n.SOC., 7 [2] 52-60 1954).
John Mandel, Statistical Analysis of Experimental Data.Inierscience Publisp ers , New York, 1964.
B. A. Proctor, I Whitney, and J W. Johnston, “Strengthof Fused Silica,” Proc. Roy. SOC., er. A, 297 [I4511 534-571967).
J. E. Ritter , Jr . , “Fatigue Strength of Silicate Glasses,”Phys. Chem. Glasses, 11 [l] 16-17 1970).