Brittle fracture in silicon studied by an hybrid Quantum/Classical method (LOTF)


T. Albaret


Brittle fracture is a nice example

of intrinsically multi-level problems in material science.

On the large scale one has to correctly take into account the

driving force of the crack that depends on the elastic

properties of the whole system. But simultaneously it is also

necessary to accurately describe the bond disruption processes that

take place at the atomic scale.

In strongly covalent materials the quantum precision

in the vicinity of the crack tip is particularly crucial.

In the case of diamond silicon

the use of purely classical potentials leads to qualitatively

wrong descriptions of the crack dynamics showing

for example  plastic deformations, high fracture

energy release rates, unphysical crack tip structures,

large surface roughness

and even crack blunting and wrong crack propagation direction.

On the other hand full quantum calculations remain practically

untractable on the several thousands of atoms needed

to integrate the slowly decaying crack elastic field.

To face this problem we used the hybrid method LOTF in

which the parameters of an adaptable potential vary

with the time to capture the quantum forces evaluated

in some selected regions of the system.

The method was applied to (110)[1-10] and

(111)[1-10] cracks upon mode I opening, each simulation

box containing around 180 000 atoms.

The results evidence

that the local accurate treatment in the crack tip region

largely influences the structure and roughness

of the opened surfaces as well as the fracture propagation direction.