Thats true, but the issue in a cubesat is going to be all about total propellant mass fraction (The fraction of the vehicle mass at launch made of of stuff you can sling out the back at high speed), so while Xe is a better reaction mass if you have the space for the tank, it may well be that in this particular use case the higher storage density (and thus the ability to fit more of it into a tiny tank) actually trumps the heavier ion.

Space propulsion is all about propellant mass fraction and exhaust velocity, as those two numbers define how much delta V you can get out of your available fuel.

The problem with light ions in this situation is that the momentum transferred is simply the product of exhaust mass and exhaust velocity, the energy required to produce that exhaust velocity is 1/2 mv^2, thus a heavier ion travelling more slowly requires less energy input to the accelerator for a given amount of momentum transfer then a light ion moving fast.

However if you have surplus electrical power, and are not too concerned about producing large accelerations (even by ion drive standards), and can solve the corrosion and thermal management problems, it might actually be a reasonable tradeoff.

All space propulsion is tradeoffs between energy/reaction mass/specific impulse/acceleration, there are no really right answers here, and having another validated tool in the box is always going to be useful.

It looks to me more likely the problem was excessive weight at the bow and stern rather then midships, the effect is called hogging and is a known way to snap a container ship (or oil tanker) in half, both have occured in the past.
Basically the keel (The BIG beam running all the way from bow to stern down the bottom of the hull) can only take so much sheer stress and if the weight distribution does not match the localised boyancy implied by the current displacement you can very easily bend the ship.

If and how it came to be loaded that way will be one of the things on the investigators list.

There is of course software used to look at this stuff but it cannot realistically be run on the dock during a very tight turnaround, so the declared weights are used as the only data available in advance of starting loading. Not only does that mess of linear algebra have to give a fully loaded ship with the centre of mass and moment of inertia in the right regions (Important for stability and handling), it must also ensure that the total cargo mass per linear meter is roughly the same as the boyancy of that meter of wetted hull at all times during the loading.

Further shippers will sometimes pay a premium for say not having a can of high value goods put in a corner on top of a stack where it is somewhat more likely to be lost, and some of those cans may be 'reefers' (Refridgerated containers) requiring both power and ventilation to remove waste heat, the problem swiftly becomes complex, doubly so as the ports stacking order also feeds into this if you want loading to go smoothly.

A nasty accident, but nobody died, and the hull and cargo will have been insured, so a better outcome then is sometimes the case.

Hope that explains why it is not just about total weight.