I can answer questions 2 and (tentatively) 4!

When freezing samples, they are cooled rapidly to form vitreous (noncrystalline) ice. If the ice warms enough (and that temp is still well below 0°C), it can transition into a crystalline form. This makes the ice expand and become spiky, which can damage proteins and cells.

For differences in LN2 usage, not every dewar is created equal. Age, the degree of vacuum between the walls, and the distance between the inner and outer walls can substantially affect the thermal conductivity, and thus the boil-off. Differences in how they are capped (which by nature can't be vacuum-insulated) can also change their efficiency.

No problem, The gas phase avoids nitrogen liquid infiltration into our storage straws that could be imperfectly heat sealed. If only a tiny amount gets in this expands 700 times in volume to the gas phase when warmed up causing the straws to pop.

The key glass transition temperature is that of the samples, there will inevitably be some warming as the tanks are used and sample racks removed, but so long as the samples don't warm more than -130°C they won't experience any structural changes.

OFAF just saves liquid nitrogen, quite a lot is lost cooling down the pipes when doing a fill, so may as well fill all of the tanks while the supply pipe is cold.

I think there is an issue with Tank 1, perhaps the insulation is not as good as the other tanks. I am monitoring to make sure it does not get worse over time.