STRENGTH OF GLASS UNDER A PROLONGED LOAD V. A. Brest and V. I. Shelyubskii
UDC 666.11.01:620.174:620.172:666.192
The use of glass as a structural material in modern engineering requires that its strength characteristics under loads for a specified time be studied. Under a prolonged steady load in a moist atmosphere glass becomes weaker. This is called static fatigue, whose characteristic feature is that when a constant stress o above a threshold o 0 (safe stress) acts on glass the glass withstands it for an interval of time ~o and then fractures. The static fatigue of glass is due to the corrosion mechanism of crack growth in a moist atmosphere and is virtually not observed in a vacuum and in a liquid nitrogen medium [I, 2]. The time Tw for which glass withstands a given stress o is called endurance. The dependence of the endurance 9 on the stress is described by [i] log 9 =
a--~ log o
(I)
where ~ and ~ are coefficients that characterize the breaking point and static fatigue, respectively. To determine ~ and ~ we studied specimens of thermally polished (TP) of sheet and quartz (KV) glasses with different mechanical working of the surface under a static load for a short and a long time. The testing was done under normal climatic conditions at 20 • 5~ and a relative humidity of 60 • 10% on custom-built equipment by centrosymmetric and transverse bending methods (Fig. i). The stresses in the specimens were calculated from the following formulas: for centrosymmetric
bending o = A(P/h2);
for transverse bending o = 3PL/(2bh2), where A is constant for glass with a given Poisson's ratio and the dimensions chosen for the specimen, supports, and loading die: here A = 0.426 for glasses of both types: P is the load; L is the distance between supports; and b and h are the width and thickness of the specimen. The specimens for centrosymmetric bending tests were square plates measuring I00 • i00 • 5 mm. the surface of the sheet glass specimens were thermally polished and the quartz glass specimens were polished mechanically (R z = 0.05). The specimens for the transverse bending tests were bars with a rectangular cross section and measured i00 • 15 • 5 mm. The end surfaces of the sheet glass specimens were ground with M40 powder and the quartz glass specimens were polished mechanically (R z = 0.i). The glass strength was determined mainly by the presence of defects on its surface. To eliminate specimens with coarse defects before the tests we inspected them visually and the quality of the grinding and polishing of the quartz glass specimens was spot-checked by etching in a hydrofluoric acid solution. On the basis of the tests we determined the short-time strength ot; each batch had no fewer than 60 specimens. The values obtained for the strength o t and ~ and ~ (at a 0.5 breaking probability) of thermally polished and quartz glasses are given in Table i. Specimens were then tested under a prolonged static load with different levels of stress acting for a base time of 500 h. Endurance tests and tests to determine ~ and were made in accordance with the data of [4]. At comparatively low stresses some specimens
NITS.
Translated from Steklo i Keramika, No. 6, pp. 14-16, June, 1991. 0361-7610/91/0506-0247512.50
9 1992 Plenum Publishing Corporation
247
TABLE 1
ffiIStrength, OS t , MPa JId~s ICOa~~ eients Glass
Imin
ax. mean :tion
Y
L
Transverse bending TP KV
49,0194,0169,0110,0 15,019nhJ,5 51.0 107,0 76,9 15,0 7,0 20• 1.5
TP UD* KV
Centrosy~etrie 59,0 197,0 J 122,0 I 23,0 50,0 l 30,0 t 49,0 271,0 136,0
KV*
bending a6,0 20.7 19• 1.5 4,0 6,9 19• 41,0 22,3 20•
24,0 50.0 41,0 5,0 7.620• ,5
*Data from [3] for samples of sheet glass (upward drawn, UD) and quartz glass with artificially induced surface defects at a breaking probability < 0.01. The other data are for a breaking probability ~ 0.5. in the series did not break during the time ~o. Those samples were eliminated from further tests and as a result samples censored from the right were formed (All-Union State Standard GOST 11.009-79). When the endurance test results are censored on the right the values of ~ are no higher, and those of 7 are no lower, than those for complete samples. To determine ~ and 7 we made a linear regression analysis of the test results and found the relation between the stress o i and the average endurance ~aveThe relations obtained for various of the specimens are shown in Fig. 2.
fracturing
probabilities
and states of the surface
We obtained 7 = 20 • 1.5 for quartz glass with a mechanically polished surface and 7 = 19 • 1.5 for thermally polished glass with a lower chemical stability (see Table i). The long-time strength of quartz glass was studied earlier on specimens with an artifically damaged surface [3] and analysis of the results of tests on these specimens showed that 7 = 20. The value 7 = 21.5 • 1.5 for quartz optical fibers in [5]. Comparison of the values of 7 in Table 1 confirms that 7 is constant for the given type of glass and does not depend on the testing method and the defect density of the surface, i.e., is determined by the chemical composition and the structure of the glass [i, 2]. The value of a depends primarily on the state of the glass surface. We must point out that ~ can be calculated not only on the basis of the endurance tests [4] but also from the results of determinations of the short-time strength Os.t. on a testing machine. Indeed for an endurance ~ = 1 sec we have from Eq. (i) = 7
Using Baily's
principle
log
of linear summation
o.
of damage,
we can easily show that at 7 =
15-30 ast ~ a~=l, where
oz=~ is the strength
Our calculations therefore, we have
at an endurance
showed that Os.t.
9 = 1 sec.
differs
=
~
from oz= I by less than 10%.
log
o t.
To within
10%,
(2)
Thus a for glasses for glasses of a known composition (known 7) with a given surface state (ground, polished, etc.) can be calculated rapidly from the results of determination of Ost.
248
a)
MPa
b~ 4 / 2
//L
-
_1
0
Fig. 1
J
12 log ~,sec Fig. 2
Fig. I. Loading devices for determining the endurance of glass under centrosymmetric (a) and transverse (b) bending, i, i') Circular and lateral supports; 2) specimen; 3, 3') circular and lateral dies; 4) force frame; 5) dynamometer; 6) loading screw; 7) weight. Fig. 2. Strength o of glass on the load time ~. i, 3) Quartz glass, fracturing probability 0.5 and <0.01, respectively; 2, 4) thermal polishing of sheet glass, fracturing probability 0.5 and 0.01, respectively; 5, 6) quartz glass and upward drawn sheet glass with artificially damaged surfaces, fracturing probability <0.01. In order to obtain the equations for endurance with a given fracturing probability we must calculate the short-time strength (see Table 1 and Fig. 2) and then determine a from Eq. ( 2 ) . The endurance equations for a fracturing probability of 0.i (~ and 7 were calculated from the data of Fig. 2) of thermally polished sheet glass and mechanically polished quartz glass, respectively, have the form: for centrosymmetric
bending log ~ = 11.0--19 log o; log 9 = 12.0--20 log o;
for transverse bending
(see Fig. 2, curves 3 and 4) log r = 11.2--19 log o; log 9 = 12.3--20 log o.
When we consider the error in the determination of 7 and the instability of the glass surface, we obtain a = ii.0 and 7 = 20 from the endurance calculation for products with an undamaged surface for thermally polished sheet glass and mechanically polished (R z = 0.I) quartz glass. The endurance equation at a fracturing probability <0o01 will have the form log T = 11.0--20 log o. We must point out that a is constant only if the state of the surface of the product remains unchanged during use. This is the case, e.g., in multicomponent designs in which other transparent components protect the surface of the load-bearing glass against damage. If the surface is damaged during use, a and, hence, the endurance decrease substantially. The minimum values in this case correspond to those observed for specimens of glass with artificially induced surface defects [3]. The lower limit of the endurance of glasses can be evaluated from the following equations (see Table 1 and Fig. 2, curves 5 and 6): for quartz glass log T = 7.6--28 log o;
249
for sheet glass log T = 6.9--19 log o. In summary, the above results make it possible to evaluate the endurance of sheet glass and quartz glass with different surface states at a given fracturing probability. LITERATURE CITED l 9
2. 3. 4. 5.
G. M. Bartenev, Mechanical Properties and Heat Treatment of Glass [in Russian], Gosstroiizdat, Moscow (1966). V. A. Bershtein, Mechanohydrolytic Processes and the Strength of Solids [in Russian], Nauka, Leningrad (1987). V. I. Shelyubskii and N. A. Velikovskaya, "Choosing the strength of glass in the design of glass products," Steklo Keram., No. 5, 4 (1971). M. N. Stepnov, statistical Methods of Processing the Results of Mechanical Tests [in Russian], Mashinostroenie, Moscow (1985). V. A. Bogatyrev, M. M. Burnov, N. N. Bechkanov, et al., "Strength of long glass optical fibers," Tr. Inst. IOFAN, 5, 60 (1987) (Institute of General Physics, Academy of Sciences of the USSR).
ANTICORROSION COATING FOR QUARTZ GLASS UDC 666.192.056:620.19
P. I. Andrienko, A. G. Varlamov, and Yu. M. Grigorev
Our goal here is to develop an anticorrosion coating for quartz glass, i.e., to provide high-temperature corrosion protection for elements of optical systems and fibers in moist oxidizing media. Several known types of high-temperature coatings for quartz glass are based on a ceramic made of A1203, Si3N~--SiC, formed by chemical vapor deposition (CVD) (U.S. Patent 4540601, Japanese patent application 63--85023) [i]. An AI203 coating adheres well to quartz glass but is stable only to 500-600~ in moist oxidizing media. Ceramic coatings of Si3N 4 and Si3N4--SiC are more resistant but are difficult to obtain in practice because these compounds are formed slowly in CVD processes. Hirai and Goto [2] reported the development of a ceramic made of amorphous silicon carbonitride with variable composition, which forms at acceptable deposition rates (up to 0.3 ~m/sec) and are distinguished by superior anticorrosion properties. By varying the elemental composition of a compound, we can vary the physicomechanical and electrophysical properties of films of this compound over a wide range and thus select the optimum coating compositions with regard to the linear thermal expansion of the glass. The latter is done here for KV quartz glass. The electrothermographic
method was used to deposit the coating.
Capillaries of KV quartz glass were pulled tightly onto a tungsten filament, which was heated electrically in a reaction gas mixture according to a program. An optical method employing an digital electrothermograph based on a MEKS minicomputer was used to the variation of the sample temperature and to maintain it. A filament with a capillary (outside diameter 350-370 ~m) was heated by an electric current to the desired temperature in a mixture of tetramethylsilane, hydrasine, and hydrogen and was maintained at that temperature to form amorphous silicon carbonitride layers on it with a thickness of 0.i20 um. The studies were carried out at 1270-1520 K~ The elemental the reagents. The electron microscope was found by local ogy of the coating
composition of the coating was varied by varying the concentration of thickness of the layers formed was determined with an MREM-100 scanning on cleavages of capillaries and the elemental composition of the coating x-ray spectral analysis on a USKhA-733 analyzer. We studied the morpholformed and its anticorrosion capability in water vapor at 920-1420 K. The
Institute of Structural Macrokinetics, Academy of Sciences of the USSR. from Steklo i Keramika, No. 6, pp. 11-17, June, 1991. 250
0361-7610/91/0506-02505i2.50
Translated
9 1992 Plenum Publishing Corporation