Okay, so the title is a bit off. They’re hunting for partial Dyson spheres using infrared and optical.

I was confused on how they would detect something completely blocking a sun from millions of light-years away.

Even a Dyson sphere, which is technically unlikely anyway, would be possible to spot. You would look for something very bright in the infrared spectrum with almost no light in the visible spectrum. It would also be larger than a normal star of the same energy, but that would be hard to tell given all the other issues.

A partial swarm is easier because it will have variability towards more infrared and then back to a more normal spectrum.

And, of course, all this is speculation until we find a candidate and determine it doesn’t have a natural source for that behavior.

Why would there necessarily be strong infrared emissions? Since a Dyson Sphere is meant to harvest all energy produced by a star, any leakage would be unnecessary inefficiency, wouldn’t it?

Because all that energy contains heat as well, and you’ll need to balance the heat from your star along with the energy absorbed.

You’re never going to get to 100% efficient conversion, so you’ll have to radiate away the heat so your sphere doesn’t melt or something.

Sure, you won’t reach 100%. But say you reach 99.9% - the Dyson sphere should radiate infrared at 0.1% of a normal star, right? It wouldn’t necessarily be bright.