Radiative heating and cooling with spectrally selective surfaces.

Matter continuously exchanges energy with its surroundings. This exchange can be dominated by radiation, conduction, or convection. In this brief review we discuss how proper design of radiative surface properties can be used for heating and cooling purposes. The desired properties can be understood once it is realized that solar and terrestrial radiation take place in different wavelength ranges and that only part of the solar spectrum is useful for vision and for photosynthesis in plants. These facts allow the possibility of tailoring the spectral absorptance, emittance, reflectance, and transmittance of a surface to meet different demands in different wavelength intervals, i.e., to take advantage of spectral selectivity. One example is the selective surface for efficient photothermal conversion of solar energy, which has high absorptance over the solar spectrum but low emittance for the longer wavelengths relevant to thermal reradiation. Below we discuss the pertinent spectral radiative properties of our ambience. These data are then used as background to the subsequent sections treating four examples of spectrally selective surfaces. The first example is the previously mentioned selective surface for converting solar radiation to useful heat. The second example considers surfaces capable of reaching low temperatures by benefiting from the spectral emittance of the clear night sky. The third example concerns two related types of transparent heat mirror. The fourth example, finally, treats radiative cooling of green leaves; this part is included since it gives a nice example of how nature solves a difficult problem in an elegant and efficient way. This example hence provides an interesting background to the other cruder types of artificial selective surfaces. Throughout our discussion we treat the ideal spectral properties, give an illustrative experimental example of how well this goal can be realized, and-where this is possible-show a corresponding theoretical curve indicating to what extent the measured results can be theoretically understood.