We should like to comment on the paper by Cooke, Irving, Booth and Beevers [l] describing observations of intergranular separation during fatigue of a medium carbon, low-alloy steel, and also on the discussion which has been generated .
The observation[l] that intergranular separation is almost absent in tests conducted in vacuum suggests that the mechanism is principally air/water-vapour stress-corrosion, and this has encouraged the present authors to conduct tests designed to investigate the effect of growth rate on the proportion of intergranular facetting observed. Specifically, referring to Fig. 1 (after Ref. [l]), we have examined the possiJbility that the decrease in the proportion of intergranular separation (%1/G) observed at high values of AK (>18 MPavm) might be a direct result of the high fatigue-crack growth rate with respect to time (daldt) which such high AK values would produce in tests at the frequency used (175 Hz).
Our initial results appear to be inconsistent with those reported by Cooke et al.
[l], and consistent with preliminary observations by Ritchie[2], in that they lead us to believe that %1/G is controlled by AK only through its effect on da/dt, and that it is also affected by K,... Cooke et al. suggest that K,,, is not a controlling parameter for %1/G. We do not, however, suggest that the work in the original paper is in error, in view of the fact that the two studies were conducted on steels of different composition, and at different frequencies. Some of the results described below show that the effect of frequency is of major importance.
The steel used was a high-strength quenched and tempered low-alloy steel, conforming to the specification En30B. Compositional and other data are given in Table . Tests were conducted in 3-point bend under constant-K conditions, using a miniature analogue computer to control automatically the loads applied by a servo-hydraulic testing machine, in response to the output of an electrical-potential crack monitoring system. This control system has been described elsewhere , and provides K control to within 3%. The environment was laboratory air, and the frequencies used were generally in the range l-200 Hz. Values of %1/G were obtained from S.E.M. photographs of the fracture surfaces. No estimates of crack growth rates were made.