when growing AlGaN on GaN template, cracking always happens. do you have any idea to reduce cracking? i am trying to use the voids inside the GaN template to relieve the strian, anybody has ideas about this method or could you give me some suggestions? many thanks.
Hi Zhi Li,
Hae-Kon is right, grading the alloy content should help but will still result in cracking eventually. See this paper for how the MOVPE gas flow cooling environment and layer properties affect AlGaN HEMT barrier layer morphology and electrical properties. The information should be useful even for thicker AlGaN layers.
https://www.researchgate.net/publication/258809770_Tensile_strain-induced_formation_of_micro-cracks_for_AlGaNGaN_heterostructures
Good luck.
Article Tensile strain-induced formation of micro-cracks for AlGaN/G...
In my opinion, atmosphere gases in cooling process are not related to micro-cracks. Surface is very impotant in HEMT process, so the right gas is used for anti-oxidation. If you want to confirm this, you can test by using the blocking layer(SiO2, SiN layer). Surface etching isn't occured, so you can ignore the cooling gas.
Hae-Kon, I find it useful to distinguish between the two types of cracking that can occur in AlGaN on GaN epilayers:
- A micro-scale crack is likely to form early in the growth process, possibly in the GaN during nucleation and possibly at the AlGaN/GaN interface, depending on the AlGaN thickness and lattice mismatch. These cracks will be visible by optical microscopy and should network along lattice planes in a hexagonal formation.
- Nano-scale cracking can occur in HEMT layers, only visibly by AFM. These cracks are due to the alleviation of strain as evidenced by their triangular morphology and prevalence at high Al content, but as you point out are only present under certain cooling conditions and may be prevented by altering the gas type. And from Zhi's description of the problem I agree it is likely to be the first case.
Thank all of you for your discussions. I think the cracks I mentioned are micro-scale cracks. The GaN template I use is ELOG-GaN, so there are many voids in the gap of patterns. And I will grow AlGaN on this template with voids insides. Do you think these voids can help me to relieve some strain?
The epilayer overgrowth should act to reduce the dislocation density in the GaN by facilitating annihilation. Whether the ELOG/voids themselves are responsible for the cracks is not clear - did you check the overgrown GaN surface for cracks also? If the ELOG-GaN surface has no cracks then the cracking is a result of strain relaxation in the AlGaN, and in this case the cracking should result in a more relaxed layer.
- Badges
- Science topic