Arc-furnaces were used for ore smelting in Sweden, Germany and the USA, more than 100 years ago.

The process still needed a reducing-agent and flux. My reference book concentrates on the electrical engineering and does not identify those two reagents, but I guess they were still coke and limestone.

This process demands vast quantities of electricity, but so would electrolysing water to give hydrogen, and it is still necessary to heat the ore.

There was very good reason for using coke, rather than coal in a blast furnace.

The charge is mixed at the top of the furnace and heats as it descends, burning at the hottest part of the ore reduction process.  Any volatiles in the coal would be either volatilised, in the absence of free oxygen, or burned in a surplus of oxygen along with the rest of the coal – in the wrong place in the furnace.  Simple as that.  Coal did not work but coke did.

Clearly, plastics could only be used by ‘firing’ at the point of reduction.  With an arc furnace, there would only be the plastic, plus whatever it is carried into the reactor by, and still the slag to be dealt with.

As soon as oxygen (even from the PET) is available the graphite electrodes may well be rapidly eroded.

Hydrogen, I suspect, is likely far better than plastics, as a reducing agent – because the process control would be protecting the electrodes.