After some more testing of my own, I realized there may be something else at work here that I've seen used in other contexts before: The arrows are governed by anchors /placement ids that go into the chemical equations themselves. The latex process needs to be run more than once: Once to establish the position of the anchors in the document, and again to draw them in the correct place. I may try to elevate this to chemfig; except for the fact that I'm just about done with my Organic Chemistry class...
chemfig drawing renders incorrectly