Simple, easy-to-use and physically meaningful analytical models have been developed for the assessment of the
combined effect of the lattice and thermal mismatch on the induced stresses in an elongated bi-material assembly, as
well as on the thermal mismatch on the thermal stresses in a tri-material assembly, in which the lattice mismatched
stresses are eliminated in one way or another. This could be done, e.g., by using a polished or an etched substrate. The
analysis is carried out in application to Gallium Nitride (GaN)-Silicon Carbide (SiC) and GaN-diamond (C) filmsubstrate
assemblies. The calculated data are obtained, assuming that no annealing or other stress reduction means is
applied. The data agree reasonably well with the reported (available) in-situ measurements. The most important
conclusion from the computed data is that even if a reasonably good lattice match takes place (as, e.g., in the case of a
GaN film fabricated on a SiC substrate, when the mismatch strain is only about 3%) and, in addition, the temperature
change (from the fabrication/growth temperature to the operation temperature) is significant (as high as 1000 ̊C), the
thermal stresses are still considerably lower than the lattice-mismatch stresses. Although there are structural and
technological means for further reduction of the lattice-mismatch stresses (e.g., by high temperature annealing or by
providing one or more buffering layers, or by using patterned or porous substrates), there is still a strong incentive to
eliminate completely the lattice mismatch stresses. This seems to be indeed possible, if polished or otherwise flattened
(e.g., chemically etched) substrates and sputter deposited GaN film is employed. In such a case only thermal stresses
remain, but even these could be reduced, if necessary, by using compliant buffering layers, including layers of variable
compliance, or by introducing variable compliance into the properly engineered substrate. In any event, it is expected
that strong adhesion could be achieved by using an appropriate fabrication technology, so that no GaN film cracking
would be possible, if the film is in tension, or delamination buckling could occur if the film is in compression. The
developed models can be used to assess the possibilities and opportunities associated with GaN materials technology.
|