Sanity check in thermal stress simulation gone wrong

I'm trying to familiarize myself with simulating thermal stresses using static studies in Solidworks Simulation. I've set up a window (fused silica) in a stainless flange that are connected rigidly:

As a sanity check I tried to simulate the stress for zero temperature difference (using soft spring stabilization). Of course I'd expect no stress whatsoever, but I get quite a sizable von Mises stress:

Oddly enough, if I isolate the flange the same thing happens. I'm guessing I must have forgotten some constraint or setting somewhere, but I haven't been able to figure it out. If anyone has any ideas that would be fantastic (I've attached my assembly file).

SolidworksSimulation