Holy moly, seriously? So the arc is actually traveling from the bottom of the button OUTSIDE the mod, and to the top of the battery?
Not really... The rhodium pin contacting the battery isn't the only moving part in play. The current moves from the neg end of the battery to the rhodium pin, the main button section, to the button collar, to the tube.
If there is any resistance in any of these connection it can result in heat or arcing. Two of these connections are made with a threaded interface (rhodium contact screws into the main button and button collar screws into the main tube). As long as those two are clean, they shouldn't pose any issues.
That leaves the interfaces between the battery and the rhodium contact and the main button and the button collar.
The battery/rhodium contact shouldn't be a problem unless there is dirt or oil coming between them preventing a solid connection. Beyond that, the fact that the contact is pressed directly against the bottom of the battery is enough to assure solid contact...
Leaving the "usual suspect"... The connection between the button and the button collar.
To look at it, it looks like there would be all kinds of surface area to make a solid electrical connection but, in reality, the actual contact area is very small. It's basically a cylinder going through a hole that slightly wider than the cylinder... Or,looking at it straight on, a circle inside a slightly wider circle.
When you do this, the two circles can only make contact at a single, tiny point.
It's very easy to move or break that point during the button press. It's in the micro second that the two circles break contact but are still a thousands of a millimeter close to each other that the arc occurs.
To make matters worse, the arc leaves behind carbon residue which in turn leads to more arcing.
The solution is the clean the button and inside of the collar really well to remove any existing carbon residue and, when firing, give a slight lateral push to force the button into the side of the collar.