I'm not sure where I might have read it, so the possibility of the ER having achieved the extra range by using a higher nickel formula may well be wrong - it's just that it seemed to be a not too unreasonable conclusion.
Maybe with a bit of maths it can be worked out? e.g. by looking at greater weight of ER over LR - which ev-database says is 100kg.
Assuming energy densities of 180Wh/kg for NMC523, 200 for NMC622 and 240 for NMC811, then if just the chemistry was changed, the capacity would increase to 71kWh for NMC622 or 85kWh with NMC811. So, it's safe to say the ER doesn't use NMC811.
Doing the calculations the other way around. 64000 Wh / 180 Wh/kg = 355Kg weight of cells for the LR. So, if the ER's extra weight is wholly down to the cells then 355kg + 100kg = 455kg x 180Wh/kg = 82kWh, which is a bit too much, but close the the ER's 77kWh.
But if, let's say, the cells are only 80kg heavier and the more powerful motor is 20kg heavier (I'm not aware of any other significant differences), then the calculation becomes 355kg + 80kg = 435kg x 180Wh/kg = 78.3kWh.
Hence my best guess would be that the chemistry is the same in both variants and about 75kg of the extra 100kg for the ER is down to more NMC523 cells.