I knew that the base-emitter junction is slowly destroyed by reverse avalanche breakdown and the internal heat produced but I did not know it is caused by bits of the junction becoming red hot.
When I worked for Philips I saw their first LED and I joked that it was red hot inside it.
On my first MOSFET driver design I put an E-B diode
where it should have been a C-B junction. Watching
through the probe station microscope I saw a flash
of light on every (failing) part and felt "stuff" hitting
my face. The heat was launching the silicon "guts"
out of their dielectric isolation tubs after reaching
orange-hot.