Hope attached articles would be of help. Best wishes, Amitava De (www.me.iitb.ac.in/~amit.html)

Also you may see our paper in which we show that there is adhesion and diffusion between aluminum alloy and steel tool

https://www.researchgate.net/publication/263163070_A_proposed_diffusion-controlled_wear_mechanism_of_alloy_steel_friction_stir_welding_%28FSW%29_tools_used_on_an_aluminum_alloy

Article A proposed diffusion-controlled wear mechanism of alloy stee...

If there is no tool wear, it remains polygonal and works fine. If you want  to remove adhered aluminum then dip the tool in NaOH solution, the aluminum dissolve in the solution.

Off course, dipping the tool in NaOH or other similar solution may dissolve the adhered material but that does not guarantee stopping of wear. Secondly, we must remember that the joining process should be able to perform long weld seam in an uniform manner without premature wear of tool profile. So, a better design of tool pin profile considering its load-bearing ability under the severe thermomechanical environment will remain to be the winner. To the best of my knowledge, most commercially successful FSW process is still conducted using a pin with circular profile only.

The case is not about the tool wear but the work material got deposited around the polygonal pin and made the profile circular after the process. what COULD BE THE REASON????  is it the process parameters isue????

For explanation of this issue it seems that the excessive heat in the weld zone is the main reason for the malfunction of tool pin profile. This excessive heat can be resulted from high rotational speed and/or low traverse speed. Threaded cylindrical and threaded tapered pins are recommended for tool pin profile. In the cases that it is impossible to change tool pin profile, reducing rotational speed could be helpful. But it should be noted that the ratio of ν/ω < 0.1 always must be taken care of.

Yours sincerely….