As explained before - Al has a high tendency for binding to oxygen. Even at a relatively high vacuum, the Al can react with the residual oxygen.

What gas purity do you use? What is the base pressure of the deposition?

One way to reduce the oxide content in the film is to reduce the O content in the chamber. Some ideas for doing that include:

Purging the gas lines before deposition (i.e. by flowing the gases for a few minutes without turning the plasma on)

Using higher purity Ar and N gases

Leak testing the gas lines and the vacuum chamber

General vacuum system maintenance such as pumps maintenance (tip seal replacement for roughing pumps, regeneration of cryo-pumps if applicable etc), baking-out the vacuum chamber, etc

In addition, although not entirely related, I've previously found that during e-beam deposition of Ta (which is also reactive with O) there is a significant effect of the deposition rate on the oxygen content. It isn't a straight forward dependence and there appears to be an optimal rate (at least for the system I've been using). You might want to try several rates in your system as well.

Good luck,

Lior

Mr.Soumendu Sinha,

Aluminium is very reactive in presence of Oxygen. You are getting this Al2O3 formation from the Oxygen content in your protective atmosphere mixture. What is the percentage of Oxygen in this mixture?

We are working on the synthesis of TiO2 layers on Aluminium. We want it straight on Al pure sheets in order to understand the photocatalytic behaviour of TiO2 on the metallic surfaces. However we clean the surface of Al with desmutting and acid cleaning, it appears that the Oxygen in the chamber is sufficient to form an Oxide layer first on Al before Ti reacts with Oxygen to form TiO2. Hence, Al2O3 is a kind of inherent with Al in presence of Oxygen. You may find our paper attached with this mail interesting.

Also, why we don't see N in your reactive sputtering. No guesses, apart from the method that you are using to find the N. Look for AlN instead in TEM!!

Dear Sir,

Thank you for sharing your experience. When we try to get AlN, for one of the samples, the atomic weight ratio obtained by EDS is

Element Atomic%

N 37.43

O 16.19

Al 45.58

We want to reduce the oxygen content to get proper AlN.

If you want to reduce the oxygen content, you can try to reduce the discharge pressure and to increase the discharge power.

Take a look to your magnetic configuration.

Thank you Sir. These are excellent articles on the topic. I will go through them.

As explained before - Al has a high tendency for binding to oxygen. Even at a relatively high vacuum, the Al can react with the residual oxygen.

What gas purity do you use? What is the base pressure of the deposition?

One way to reduce the oxide content in the film is to reduce the O content in the chamber. Some ideas for doing that include:

Purging the gas lines before deposition (i.e. by flowing the gases for a few minutes without turning the plasma on)

Using higher purity Ar and N gases

Leak testing the gas lines and the vacuum chamber

General vacuum system maintenance such as pumps maintenance (tip seal replacement for roughing pumps, regeneration of cryo-pumps if applicable etc), baking-out the vacuum chamber, etc

In addition, although not entirely related, I've previously found that during e-beam deposition of Ta (which is also reactive with O) there is a significant effect of the deposition rate on the oxygen content. It isn't a straight forward dependence and there appears to be an optimal rate (at least for the system I've been using). You might want to try several rates in your system as well.

Good luck,

Lior

I completely agree with Lior Kornblum, this is may be cross-contamination issue, which might be due to using the same chamber for Oxide and Nitride deposition.

Cleaning the gas lines with purging Ar or N should definitely help.

Dear Soumendu,

In contineuation to all the superb answers above the main culprit of oxygen contamination is from the Nitrogen and Argon gas cylinder. you may try to put oxygen traps in the gas lines. Also try to avoid using the gas cylinders which are nearing the empty mark.

This remark is with the assumption that your chamber and gas lines do not have leak.

Thank you everyone for the excellent replies. It is extremely difficult to overcome the oxygen residuals even at very high vacuum (1E-7 Torr base pressure). We are also aware of the contamination that may come through the gas lines/ cylinders . By taking into account the suggestions, we are now able to reduce the oxygen content to less than 5%, which is remarkably close to the samples shared in the article by Dr. P.Y.Jouan.

The incorporation of oxygen in films can attributed to impurity of injected nitrogen gas or the residue gas after pumping or the exposition the films in air after deposition or the residue superior oxide layer substrate

The base pressure is not enough, when the sputtering aerobic participation. Characterization of the sample before reaction with the morphology of the oxygen in the air.

Can you measure the amount of oxygen present in the sputtering chamber during the nitride deposition by using a gas analizer? If you are sure that no oxygen is introduced from the gas line and the vacuum limit is sufficient, you may solve your problem by making a strong pre-sputtering of the Al target using only Ar. The surface of the target is certainly contaminated by oxygen, forming Al2O3. This is difficult to remove completely if you are using Ar + N2. Conclusing: do a presputtering in pure Ar observing the deposition rate, when you pass from the "insulating"regime to the "metallic" regime, the deposition rate increases by 10-20 times; at this point the surface of the target should be ready to perform the deposition of the nitride. Good work!

Are you doing the EDS analysis without breaking the vacuum or is your sample exposed to air?

Our EDS facility is set up elsewhere in the lab. We need to take out the sample from sputtering chamber for EDS analysis, exposing it to outside air.

Either you catch the oxygen during the process in the vacuum chamber or the sample is oxidizing in the air during the time of transfer. If your sample is above room temperature, the oxidization might even be faster (it is the case in our lab with molybdenum samples).

Residual water vapor is a notorious contributor of oxygen and hydrogen, even given your low base pressure. Conditioning by prebaking the system and substrate prior to use is mandatory. Baking during pumpdown and deposition is also helpful as is long exposure to vacuum before initiating the deposition. When the magnetron is lit, further desorption of water vapor from the chamber walls and hardware will occur due to excitation by UV and particle bombardment. Therefore, it is advisable to run the magnetron for some time before opening the shutter to expose the substrate.

I see 1e-7 torr somewhere here. My experience with ECR sources would indicate at this pressure range you still may see oxygen outgassing from the walls of the vacuum vessel for 1-2 days into the running.  Are you purging dry nitrogen in your installation when you are opening to "air" for loading/unloading ?