[Enginerrrd]

>Now, I don't know if this exists but it would be nice if there was a kind of glass that let radiation pass only to one side and not the other way, a kind of "valve", this could solve the problem of water frozing from radiation in our solar heating.

This doesn't exist and can't even in principle because it would be a fundamental violation of thermodynamics. Basically, it would be a Maxwell's demon for radiation that would allow you to arbitrarily reduce entropy.

[KMag]

But, it might be possible to have a material that has different characteristics at different temperatures, as long as it's symmetric. If it's nearly opaque to IR in its cold state, hopefully sunlight at dawn would warm it rapidly enough that it would automatically "shut off" on cold nights, and still "turn on" shortly after dawn even on cold days.

[Enginerrrd]

Yes I think that'd be ok. You'd basically just be taking advantage of the thermal gradient between the radiating body and the object being heated. It'd be analogous to adding more insulation as it heats up and removing insulation as it cools down.

You'd be slowing the flow of heat into your reservoir just as much as you'd be slowing down the loss of that heat later though.

You might still get some benefit in preventing freezing at the expense of needed a larger area to get the same amount of heat flow in to your water system though.

[sgc]

Light meter controlled / tracking louvers would be the quick fix.

[kbutler]

We call this a one-way mirror.

https://en.m.wikipedia.org/wiki/One-way_mirror

They are not perfect (intensity & wavelength limitations), but they exist.

See also mirages and total internal reflection.

[K0balt]

I don’t think maxwells demon is impossible, it just is a device for converting información to energy. In theory, many such devices may be possible for the conversion of information to both energy and matter. This seems like it might be support the simulation hypothesis.