I don't know how much an ion engine costs. I imagine it isn't very much compared to the mission cost. Being a very limited production item, the per unit costs could be significant. However, even spacecraft with ion engines still need hypergolic-fueled thrusters for things like possible large course corrections and reaction control wheel unloads, so adding an ion engine does add some cost compared to not having one. There's also the additional tanks and plumbing for the xenon propellant but again, that isn't a significant cost.

The increased electrical loads that the ion engine puts on the mission mean you need a more expensive power system. This is where most of the extra cost comes in. For example, on a PKE type mission, your basic space probe might only need 100-200 watts or so to power the mission instruments. However, the ion engine itself requires a lot more power. Like I mentioned in a previous post, the ion engine on the DS-1 required about 2.5 kilowatts to reach full power. RTGs or SRTGs aren't cheap by a longshot, largely due to the costs of plutonium and the extreme safety margins built into the systems.

DS-1 proved several interesting technologies in addition to the ion propulsion. One of the most interesting was the self-navigation system. Navigation under a constant low thrust is different than conventional interplanetary navigation. DS-1 proved it's possible for the vehicle to navigate itself. In addition, DS-1 proved a vehicle can diagnose system malfunctions and take corrective actions on its own. That would also be important on a mission lasting years where radio propagation time between the vehicle and Earth is many hours.

A really long mission like PKE could possibly be shortened by a few years with ion propulsion. There would be some savings as a result of reduced Deep Space Network support over the mission as well as by not needing to maintain the mission operations team for so many years. Adding in those DS-1 technologies would further help reduce DSN and mission costs.