There are quite a few possibilities here. Obviously, surface contamination could be one cause of delamination.

I haven't sputtered graphite before, but have deposited a variety of DLC-type multi- and single-layers on metallic substrates, including 316SS, using PECVD.

For good film adhesion, I found that surface preparation was really important; ultrasonic cleaning as Polyana suggested, followed by an optimised argon etch as Maik suggested, and/or a substrate pre-heating step are all processes that might help you.

It's possible also that your films are inherently high in compressive stress, which will limit adhesion and cohesion, and also limit the thickness of film you can deposit.

What thickness of film are you depositing?

Dependent on your application, it might also be beneficial to deposit a kind of adhesion-promoting layer prior to depositing the carbon film.

To clean the surface you can immerse it on Extran 10% solution and let it in ultra-sound bath for 15 min. Repeat this procedure using destiled water and finally pure aceton. The delamination of your film is localized or it is complete?

Do you etch the surface by ion bombardement before coating deposition? How thick are the films? Do you use bias voltage (value)?

There are quite a few possibilities here. Obviously, surface contamination could be one cause of delamination.

I haven't sputtered graphite before, but have deposited a variety of DLC-type multi- and single-layers on metallic substrates, including 316SS, using PECVD.

For good film adhesion, I found that surface preparation was really important; ultrasonic cleaning as Polyana suggested, followed by an optimised argon etch as Maik suggested, and/or a substrate pre-heating step are all processes that might help you.

It's possible also that your films are inherently high in compressive stress, which will limit adhesion and cohesion, and also limit the thickness of film you can deposit.

What thickness of film are you depositing?

Dependent on your application, it might also be beneficial to deposit a kind of adhesion-promoting layer prior to depositing the carbon film.

Try to clean with Chloroform in the ultra-sound-bath for 15 min and later in the plasma cleaner for 15 min, usually it s really clean after these 2 steps.

I agree with Ross on looking at ultrasonicating, Ar plasma treatment and the possibility of interlayers. The plasma treatment does not only clean but modifies the surface of the steel (which may be what you want, or not depending on what you are using this for). We've recently investigated the effect of changing the plasma conditions on steel surface structure and film adhesion, the optimization is crucial:

https://www.researchgate.net/publication/234144688_Argon_plasma_treatment_techniques_on_steel_and_effects_on_diamond-like_carbon_structure_and_delamination

Article Argon plasma treatment techniques on steel and effects on di...

Ross gives a very good description of important steps for depositing coatings. Additionally, an important point is the use of bias voltage during the film formation. Using too high voltages leads to stress, presumably resulting in delamination.

Actually the most efficient technique for surface cleaning and get perfect non contaminated one, is obtained by Glow Discharge which is very similar to DLC mentionned by Ross Birney. the Japanse company Horiba propose a material called TENSEC which is very convenient. Other alternative is to used a GD-OES if you have one. Just install the sample on the specimen chamber and apply a soft Argon plasma (low pressure and power during 2 seconds using pulse mode). Professor Kenichi Shimidzu from Keio University (Japan) has proposed this procédure some years ago and got fantastic results on metallic surface or semi conductor. This procedure avoid chemical agents which can let frequently very thin film and ion etching which is always very delicate to operate especially for nanomaterial or non conductive ones.

If you need more details feel free to contact me.

Typically degreasing of SS136 steel parts used in UHV chambers is performed by ultrasonic baths in acetone followed by isopropanol or anhydrous ethanol. Once in vacuum the metal piece could be heated up to 180-200°C, to degas adsorbed water and residual hydrocarbons. Finally Ar ion sputtering or plasma etching should complete the task. Maybe ion sputtering would leave a very rough surface at micro/nano-scale.

Oxygen plasma cleaning followed by sputter cleaning. If you have an RIE available for not-so-nice materials like 316 run a process at 30 - 100 mtorr O2 followed by a step at the lowest possible Ar pressure.

optimizing adhesion of graphite coatings is hard ... first of all degrease by washing with acid cleaning in ultrasonic bath (commercial products are OK), carefully dry and then be sure to make a very good ion cleaning before deposition!