Why Do Voids Form at the Thin-Film Substrate Interface After Annealing?

[lamejane]

TL;DR Summary

Occurence of Voids after Annealing

Hello to all the experts out there!: )

I have encountered a problem on my samples after Post deposition annealing [ Please refer to the crossectional SEM picture - B below].

Sample description : Thinfilm metal stack [ Pd/Ti/Pd/Au] deposited onto GaAs substrate using PVD
Post deposition annealing : Using a Hotplate exposed to clean room air
--> 60 min @ 390°C [ Picture A]
--> additionally 70 min @350°C on same sample B [ Picture B]

My Question to the Experts :
I see occurences of Voids at the thinfilm -Substrate interface [ see picture B] after annealing at 350°C for 70 min . What effect could explain this phenomenon ?
Is it due to an increase in tensile stress at the interface because of the high annealing temperature ? or is there a strong solid state diffusion of the metal atoms in the thinfilm leading to defects ?

The thinfilm metal stack contains : Pd/Ti/Pd/Au/15nm/30nm/30nm/250nm and is meant to serve as an Ohmic contact

 

[Lord Jestocost]

What are the reasons for your choice of post-deposition annealing conditions (oxidizing atmosphere (air), prolonged annealing times in the range of 1 h). There could be many effects to explain the phenomenon of void forming: Oxidation, formation of new phases, Arsenic out diffusion etc.

The metal layers have thicknesses in the range of nanometers. In case of such metal-multilayer contacts, the post-deposition annealing to form good ohmic contacts is generally carried out by the rapid thermal annealing (RTA) process in a forming gas (N₂/H₂) atmosphere at temperatures between 300 ‒ 450 °C for 10 ‒100 sec.

See, for example: E. F. Chor and D. Zhang, “Electrical characterization, metallurgical investigation, and thermal stability studies of (Pd, Ti, Au)-based ohmic contacts”, Journal of Applied Physics 87, 2437 (2000)

 

[lamejane]

What are the reasons for your choice of post-deposition annealing conditions (oxidizing atmosphere (air), prolonged annealing times in the range of 1 h). There could be many effects to explain the phenomenon of void forming: Oxidation, formation of new phases, Arsenic out diffusion etc.

The metal layers have thicknesses in the range of nanometers. In case of such metal-multilayer contacts, the post-deposition annealing to form good ohmic contacts is generally carried out by the rapid thermal annealing (RTA) process in a forming gas (N₂/H₂) atmosphere at temperatures between 300 ‒ 450 °C for 10 ‒100 sec.

See, for example: E. F. Chor and D. Zhang, “Electrical characterization, metallurgical investigation, and thermal stability studies of (Pd, Ti, Au)-based ohmic contacts”, Journal of Applied Physics 87, 2437 (2000)

Thank you for your reply.
Using the Hot plate was only for test purposes. For the the long term i would move to a RTA tool as you suggested.
Lets say the process is carried out in an RTA und N2 atmosphere , what effect does a longer annealing tme ( ~ 60 min ) have on the metal layers ? Why is 10-100 sec optimal ?