Devin-M said:
The current through the external DC circuit of the 300k IR detector is directly proportional to the photon count of 3.5 micrometer photons reaching the active area of the detector and increasing the size of the sensor in your box scenario (with 300k maintained walls) will increase the photon count onto the active material in the pn junction and eventually with large enough size you get the same incident power as was used by the researchers.
https://www.speakev.com/cdn-cgi/image/format=auto,onerror=redirect,width=1920,height=1920,fit=scale-down/https://www.speakev.com/attachments/5000b9e5-011d-4ef0-af59-bedb15ff822e-jpeg.156575/
Please realize that the figure you show here from the paper was
not generated using a 300 K heat source. The photodetector was held at 300 K, not the source of the 3.5 micrometer photons.
I'm now going to come clean and give you the answer to the resizable box problem: The answer is there is nothing you can do increase the photon power flux reaching the sensor, given the restrictions of the problem. It can't be done. The power flux is dependent upon temperature, and that's it.
And you can make the box as big as you like. Put an entire ocean in the box (let's ignore the corresponding vapor pressure for the moment). One might say, "Good golly, an entire ocean, there must be a lot of photons from that!" But all those photons are spread out over a huge area. The photon power flux does not change, so long as everything is kept at 300 K, not matter how big something is. It's always the same, tiny, tiny number.
And at 300 K, the photon power flux of photons in the region of 3.5 micrometers is tiny. Really, really, tiny tiny. Furthermore, when everything is at the same temperature (glass of water + detector), there is no conceivable way to collimate these tiny, tiny amount of photons onto the detector. The power flux leaving the detector is the same as the power flux arriving at the detector, and there isn't any way to change that so long as everything involved is at the same temperature.
The figure you quoted from the paper was made with 3.5 micrometer photon power flux that I'm guessing is billions or trillions (I don't know, some huge, huge number) of times greater than that fixed, small, tiny, tiny photon power flux that could be achieved from objects at 300 K. In order to make such a power flux, a source much hotter than 300 K is needed (and yes, an IR laser counts as a much hotter source. An incandecent bulb and a diffraction grating will also suffice).
And finally, there's no point in claiming that one doesn't need billions or trillions of times greater, and a smaller number generated at 300 K will suffice. Because if the photon power flux is generated at 300 K, that's the same power flux
leaving the detector when it is at 300 K. The net power flux is zero. No power can be extracted from that.