Stacking-fault energy and yield stress asymmetry in molybdenum disilicide

Stacking-fault energies in MoSi2 due to shear along [331] have been calcula ted by ab-initio and modified embedded-atom method (MEAM) calculations. The results are used to investigate the configurations of 1/2 [331] dislocatio ns and their mobility. Shear of 1/6 [331] in the {103} plane of MoSi2 produ ces an antiphase boundary (APB) whose geometry, called APB(1), is different from that produced by 1/6 [331] in the opposite direction, APB(2). MEAM ca lculations show that APB(1) is stable while both types of calculation show that APB(2) is unstable. Both ab-initio and MEAM calculations show that the re is a stable fault close to APB(2) with a displacement of about 1/8 [331] in the same direction. The calculations also show that there is a stable f ault in the {110} plane with a displacement of 1/4 [111]. The identical fau lt is produced by a shear of 1/4 [331]. There is good agreement between the fault energies calculated by the two methods and also with the experimenta l value (200-370 mJ m(2)). The agreement between the calculated fault energ ies in the {013} plane is not so good. One factor is that the relaxation pr ocedures are different; the MEAM method has more flexibility as well as a l arger number of atoms, possibly explaining why it gives lower stable fault energies. The {103} planes have an unusual five-layer ABCDE stacking sequen ce with successive planes offset by 1/5 [301]. Shear of 1/10 [351] in the c orrect direction gives a low-energy fault with Mo atoms surrounded by the c orrect number (ten) of Si nearest neighbours. This vector is close to the 1 /8 [331] shear that produces a stable fault and may explain its low calcula ted energy. Various dissociated configurations of 1/2 [331] dislocations ar e considered on the basis of 1/6 [331], 1/8 [331], 1/4 [331] and 1/10 [351] partials. All can have asymmetrical arrangements which will respond differ ently to the direction of the applied stress, explaining why {103} [331] sl ip is much easier for crystals compressed along [100] than along [001].