Page 154 - Fiber Fracture

STRENGTH OF GLASS FIBERS 139

Extrinsic Inert Strength

Inert Strength Distributions

Inert extrinsic strength is of interest to fiber manufacturers as it is the inert strength
which is actually 'sold' by a fiber producer. Inert strength can be measured by testing
(a) at liquid nitrogen temperatures, (b) at room temperature in vacuum, (c) at room
temperature using very high strain rates, and (d) at room temperature in moisture-free
inert environment such as dry nitrogen or in oil. Measurements on small-length samples
are convenient but do not sample large flaws. The measurements of inert strengths of
long lengths of fibers are difficult but are necessary to sample the low-frequency large
flaws. These inert strengths can be estimated from measured room temperature strengths
by correcting for the fatigue effects. If the inert strength distribution is single-mode
Weibull (as given by Eq. 1) then the measured strength distribution during a constant
strain rate, ct, experiment at a temperature, T, follows:
+
[I
P(S) = 1 - exp 1 -(s/s~)~ R(S/S~)~]'~''~-*))] (18)
where
R = [(N -2)VoX"exp(-Q/RT)Y2$]/[2(Nf I)K,2Ect] (19)
Glaesemann and Helfinstine (1994) have used a similar equation to reconstruct inert
strengths of long silica fibers (lengths = 823 km) from measured strength distributions
at room temperature.

Fractography of Low-Strength Breaks

Fractography is a powerful method of identifying the nature and location of the fracture
initiating flaws (Mecholsky, 1994). It is not possible to use fractography for high-
strength fibers because, upon fracture, high-strength fibers disintegrate into powder.
When fiber strengths are low (tl GPa), it generally becomes possible to capture the
fracture surfaces and examine them by scanning electron microscopy. This requires
examination of several fracture surfaces, since most of the time no flaw is observed.
Even when one locates a flaw, there is not a single type that stands out in a population
of fracture surfaces. Sometimes, however, unexpected flaws show up. For example,
Fig. 3 shows a SEM image of an E-glass fiber fracture surface which failed at room
temperature at a strength of 745 MPa (Gupta, 1994). In this case the fiber was drawn
using a one-hole platinum bushing in the laboratory. The fracturc was clearly initiated
by the crystalline platinum inclusion in the bulk of the fiber. The precise mechanism
of the melt-platinum interaction which generated such a crystal is unclear. A possible
mechanism is oxidation of platinum as oxide vapor, followed by dissolution of some
of the oxide vapor in the melt, followed by precipitation of the platinum crystal during
cooling. An interesting issue is that the crack size estimated using the Griffith -Irwin
relation (Fq. 13) corresponding to the strength of 745 MPa is less than 0.5 Fm but the
platinum inclusion is much larger (at least 2 pm). Clearly the inclusion itself cannot be
treated as a crack. One possible explanation is that during cooling microcracks nucleate