Fellas, thank you for your interest and advice, much appreciated. The left hand configuration in Eric's diagram is pretty much the situation, although the coil/diode and S1 are interchanged.
It's part of the control setup for a PTO-driven mobile soft-serve churn in a vehicle. In fact there are two relay-driven automotive clutches. One through a bunch of pulleys and idlers drives the churn's beater blade, the condenser cooling fan, and an open-drive 1.5HP open-drive Blissfield compressor. The second clutch cuts the compressor in and out of this arrangement depending on viscosity of the soft serve in the churn's barrel. The compressor can only work when both clutches are energized, and the churn (upon product draw-off) initiates the timing sequences that result in one or both clutches turning on.
Eric, does your top graph show that in fact the max current I of L1 (the clutch coil) and therefore the diode is only 300mA? Well within the diode specs. The clutches come in and out once, twice, maybe thrice a minute but this can go on several hours if at an event where there are a lot of people present.
I was pleased to eliminate the arc but taken by surprise by the delay in the clutch(es) dropping out so it has been good learning more about back EMF.
The main thing is, can I use it with just the diode for the meantime and if the delay becomes a problem add the zener in as well.
Thank you for the suggestion.
By the way, just for clarity, it's In4004 not In4007.
Still interested to hear more discussion.
Paul