I deposited around 50 nm thick FeGe film on a silicon wafer using a DC magnetron sputtering. The compositional analysis data from FESEM EDAX at the deposition area shows that the oxygen atomic percentage of the film is nearly ten times more than from the uncoated site. What might be the reason for this difference?
One of the reasons might be the presence of oxygen in the vacuum chamber during the sputtering. What was you base pressure?
Another reason might be that the targets were oxidized, which can be usually solved by a pre-sputtering, to clean them.
Dear Sateesh,
Most probably you have been doing this deposition in a vacuum with indeed a too high base pressure. Even with a low base pressure, if you didn't use a cryopump and/or didn't have a Meissner trap in your chamber, the residual gas will have been mostly water vapour. Ge (as is the case with Si) and Fe are easily gettering oxygen or moisture. Did you do a reactive DC glow discharge pre-treatment on the wafer before sputter coating ? This might be another reason for increased pick up of oxygen by the Si-wafer (growth of the native oxide).
Dear Sateesh,
As far as I believe, the EDS is not a very accurate way to judge the elemental composition. Especially, when it comes to oxygen, carbon, or other lighter elements (with low atomic numbers), it is always problematic to quantify them using EDS. Moreover, EDS only gives the elemental composition of the spot area, not the entire sample. So, If EDS is necessary, then go for a large area scan or multiple scans and take the mean out of it.
Now let us discuss what are the factors that cause the increase in oxygen content. It can be caused by various factors.
- The very first factor is the purity of the working gas, please check what grade are you using and how much air/O2 content it carries.
- The base vacuum also plays an important role in the quality of the film. Always try to achieve the highest level of vacuum before allowing the working gas.
- Plasma cleaning is always recommended for both substrate and target material prior to deposition.
- Most importantly, If you are using a high substrate temperature (>300 Celcius), don't vent the chamber until the substrate temperature reaches 50 degrees Celsius. Venting/opening the chamber at high temperatures with air is the most vital reason for the oxidation of the thin films. So always vent with argon and open the chamber when the temperature reaches near the room temperature value.
- As far as, the difference in oxygen concentration is concerned. It may be related to the more oxidation potential of Iron (coating) as compared to Si wafers.
Sincerely...
Dear Thomas, Paul and Nisar,
Thank you so much for the replies. I am using a 2-inch target DC magnetron sputtering, which has a Diffusion pump that can create a vacuum of upto 1e-6 mbar. I am starting deposition under the base vacuum at 5e-6 mbar, with an Ar pressure of 1e-2 mbar. I am doing 3 -5 min pre-sputtering before beginning the actual deposition on the substrate.
Paul, we have a liquid nitrogen trap, but I didn't use it. Before sputter coating, I didn't do a reactive DC glow discharge pre-treatment on the wafer. Can you briefly explain how to do it?
Nisar, I am trying to prepare a thin film of FeGe with an equal composition.
Is there any other technique for the compositional analysis of thin films?
I scanned the sample at different locations using EDS and took the average, but I didn't see much difference in the oxygen atomic percentage.
We use high-grade Argon gas but don't have an oxygen trap.
I plan to prepare the films yet elevated temperatures, for optimization purposes I am doing yet room temperature. I am using N2 for venting the chamber.
Best regards,
Sateesh
Hi Sateesh, use your liquid nitrogen trap! For a DC glow discharge pre-treatment, you need a specific source. Probably not possible with your mini sputtering system (target of 2").
Are you sure you don't have a vacuum leak? Check by switching of the pumping system and keep the chamber unvented...
Hi Paul, I did external leak checking using a soap solution at the connecting valves of Argon gas flow. I found a leak at the valve connected to the argon gas regulator then I solved it. Once the rotary pump created the rough vacuum, I opened the argon gas path to the chamber using a needle valve by closing the gas at the cylinder. I have observed the steady rough vacuum condition.
If we switch off the pumping system, the rough vacuum falls slowly to 0.5 mbar within half an hour. But the chamber will not open until we vent the system with N2 gas.
Usually, my system takes 20 mins to create a rough vacuum, then I switch on the Diffusion pump for 30 min, and after that, I open the high vacuum butterfly valve. Within 10 minutes, I used to get the 5e-6 vacuum. Then I allow argon gas using a needle valve. Once the chamber pressure stabilizes, I turn on the DC power supply. Within a minute, the flickering stops and the plasma stay steady.
You have to check what happens when you have pumped down to 5.E-06 mbar and then switch off all the pumping: how quickly does your pressure rise ?
A soap test is not adequate for checking small vacuum leaks.
Hi Paul, Once the chamber pressure reached 5.E-06, I turned off the diffusion pump and kept the rotary pump in backing mode as I couldn't completely turn it off. After half an hour, the pressure rose to 0.05 mbar; after one hour, it was yet 0.1 mbar.
- Badges
- Science topic
- Similar topics
- Geoscience
- Asia
- India