How would you approach creating a material/device that lets the light through from one side, but reflects it from the other? NOT a one-way mirror that works by a different principle and doesn't fit the description. And NOT a deformable, micro-scale mirror currently used in science bearing the same name. The material/device I'm talking about should ideally have close to full permeability of light from one side and close to full reflectivity from the other.
The background and the feasibility of the idea:
I initially came up with a concept for such a material when thinking about ways to terraform Mars and make it permanently habitable. One of the main principles for doing that is heating up the Martian atmosphere to create greenhouse effect. I thought about ways to do this without introducing large amounts of greenhouse gases. I envisioned many satellites orbiting Mars (in a Starlink manner) and carrying "membrane mirrors" that would each have a huge surface area.
Using orbital mirrors for this has been proposed before, but the more orbital surface area you cover with mirrors, the less light you allow to come through at the same time, so having "membrane" mirrors would greatly increase the efficiency of such system. At first thought, this seemed like a possible way to heat the Martian surface quickly. At the second - it's mostly science fiction and might be totally infeasible. In any case, the challenge of creating such a mirror as well as the question about its feasibility remained.
People have asked similar questions on quora and other places before. Some comment on the impossibility of such a tool since it would allow violating the second law of thermodynamics. But turns out that optical isolators that allow light permeability in only one direction exist. They don't violate the laws of physics since heat (not light) is transferred backwards to an extent - electromagnetic radiation heats the "back surface" through which it can't penetrate.
So an optical isolator is a "one-way window" that lets the light through from one side and absorbs it from the other. Is it possible then to go technologically one step forwards and make the back surface not to absorb, but to reflect the light?
The reflective vs absorbent surface should increase the amount of energy that stays on the inner side of the mirror, but again shouldn't (and correct me if I'm wrong here) violate the laws of physics. The latter is because there is no such thing as a perfect mirror, meaning that even the most sophisticated mirrors have <100% reflectiveness. This means that some of the electromagnetic radiation will be absorbed and/or scattered by the back surface, which in turn means that some of the energy will be transferred back and this again won't violate the laws of physics. The question then remains - is it technologically feasible to do this and if so, how?