Page 335 - T. Anderson-Fracture Mechanics - Fundamentals and Applns.-CRC (2005)

1656_C007.fm Page 315 Monday, May 23, 2005 5:54 PM

Fracture Toughness Testing of Metals 315

FIGURE 7.18 Theoretical calculation of the P max /P Q ratio, assuming all nonlinearity is due to crack growth
and the R curve follows Equation (7.3). Taken from Wallin, K., ‘‘Critical Assessment of the Standard E 399.’’
ASTM STP 1461, American Society for Testing and Materials, Philadelphia, PA, 2004.

specimen width W. The restriction of Equation (7.2c) is satisfied only for m < 0.10. Therefore, ASTM
E 399 is applicable only to materials with relatively flat R curves, irrespective of specimen size. 2
For materials that fail by cleavage fracture, the nonlinearity in the load-displacement curve is
due to plastic zone formation rather than stable crack growth. In such cases, the 5% secant method
is totally inappropriate for estimating the fracture toughness. Rather, the load at which unstable
crack extension occurs should be used to compute the toughness. If the nonlinearity in the load-
displacement curve is small, the fracture load can be substituted into Equation (7.1) to compute
K . If there is a significant nonlinearity due to plastic deformation, the toughness should be
Q
characterized by J or CTOD. A key advantage to generalized test methods such as ASTM E 1820
and BS 7448 is that they allow for both linear elastic and elastic-plastic material behavior.
As of this writing, the ASTM Committee E08 on fatigue and fracture is considering a complete
overhaul of the E 399 standard to address the problems described earlier. Wallin [13] has outlined
a proposed framework for an improved E 399 test method. He states that K should be based on
Q
a fixed amount of absolute crack growth rather than 2% of the ligament. Given that such a

2 This analysis is based on the assumption of a power-law R curve. Real materials do not exhibit an indefinitely rising R
curve, however. When ductile crack extension occurs in small-scale yielding, the material resistance eventually reaches a
steady state, as discussed in Section 3.5.2. Therefore, it is theoretically possible to test a specimen of sufficient size to
meet the E 399 requirements, even if the initial R curve slope is relatively steep. However, since specimen size is limited
by practical considerations, the current version of E 399 is suitable only for materials with relatively flat R curves.