Its hard to predict what maximum current or voltage you can use without destroying your junctions as there are many factors associated with it. As a practical suggestion, start your measurement with lowest current (or low voltage range) and increase it in small steps so that you can have some idea when the junctions can not handle the bias stress any more.

HI Edgar, I believe atleast for one of the cases you need to carry out a breakdown voltage test using which you should be able to estimate for various sizes. Also, note that this estimation will still be very limited based on the thickness.

As Sayani indicated, it is difficult to predict max. current.

It is usual when the area is larger, the device fails early. In general, to overcome the short circuit problem or burning, you can always extend your measurement window beyond the break down voltage if you do it pulsed. The first thing is figure out what is the best pulsing conditions. You have to give enough time to the system to cool down between two pulses. I hope this helps.