I am trying to design micro-heaters that should go more than 800 deg C. I tried different materials, but tungsten seems a potential candidate. I am using lift-off to pattern the thin films of about 100-200 nm thick. I can see weird connections from corner of square pads connected to some other thin beams (literally like roots attached to corner of devices or exactly in the mid of thin beams). Looks like, stresses get developed in tungsten films.
I am using really good pressure in sputtering system and RCA1 cleaned substrates. I can see varying the deposition parameters leads to different sheet resistivity of material. What could be the main problem and how to avoid! I'll appreciate help on this question.
Tejinder, could you post a photo of your films with considered defects?
I guess that in case of stressed film it's preferably to sputter the film onto a whole wafer ant than etchdown the undesired metall. Thus avoiding a resist mask, you can heat the substrate during deposition and it probably can solve a problem. Maybe even perform a series of temperature cycling during continious deposition for stress release.
Hi Vadil'evskii,
Thank you for replying. I have attached optical photos of tungsten films showing stress, I have seen stresses getting developed even without using mask (another picture without any features present), the moment film thickness go beyond 150 nm. Stress increase with time though. I tried heating the substrate but stress tends to decrease, but can not be avoided. I'll definitely try temperature cycling. What temperature range do you suggest? May be cycling RT to 100 deg C or something else may be!!
Hi Tejinder,
The photo you attached shows classic compressive stress failure. The 'wiggly worm' or 'serpentine' is characteristic of compressive stress state in the film. The good news is that stress can be modulated in sputtered films of most transition metals. Lower pressure = compressive and higher pressure = tensile. Try increasing your deposition (not base) pressure. Each metal and alloy has its own characteristic pressure (and deposition rate) at which the film stress crosses zero. This effect was first reported (to my knowledge) in Thornton et al, Thin Solid Films vol 64, issue 1, pp 111-119 (1979). Should be many more recent citations. Another way to modulate stress, if you have it available, is to use bias sputtering (wherein a potential is applied to the substrate). RF can be used for bias sputter deposition on non-conductive substrates. Other than that your thickness range of 100 to 200 nm is good. Cycling over your desired temperature range is going to create some thermal mismatch issues that will challenge film adhesion to the substrate. Might want to consider a way to keep thermal expansion coefficient of film and substrate as similar as you can.
The abstract of the Thornton article contains quite a bit of info:
Abstract
Previous studies of the internal stresses in thin films deposited at low T/Tm using cylindrical post-cathode magnetron sputtering sources have identified a transition working pressure for each coating material, below which the internal film stresses are compressive and above which they are tensile. New stress versus working pressure measurements for gadolinium, rhodium, chromium and three nickel-based alloys substantiate the generality of the stress-pressure dependence. The occurrence of compressive stresses correlates with the presence of entrapped argon in the deposits; however, the estimated magnitudes of the corresponding compressive strains remain invariant over large changes in the argon content. It is concluded that the mere presence of argon in the films does not cause compression. When materials sputtered from both post-cathode and hollow-cathode magnetron sources were deposited onto specially shielded substrates, large compressive stresses occurred only where geometric factors indicated enhanced bombardment by energetic argon reflected from the cathode. Atomic peening (forward sputtering and recoil damage to the depositing films) is suggested as a cause of the compressive stresses in the films sputtered at low working pressures.
Hi Eric,
Thank you for the detailed description. I will go through the article and we have DC/RF bias for substrate during deposition in our systems. I am going to try that. I appreciate your valuable inputs.
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Materials Science
- Thin Films and Nanotechnology