Judging by Figure 4 there, I’d say that the walls are all nodes, which are where the microwave’s strength is zero.
It’s hard to make out what is the pan and what isn’t so I might be wrong about the boundary being all nodes. Still, there are two basic (idealized) ways waves can reflect off of a surface. In one the phases rotate 180-degrees and cancel each other out. That leads to having the boundary all nodes. In the other the wave’s phase is unchanged and you get reinforcement at the boundary, leading to the boundary being all anti-nodes. (Non-idealized settings are some mixture of these, plus absorption.) To me, it looks more consistent with the former than the latter.
