Radiative sky cooling cools an object on the earth by emitting thermal infrared radiation to the cold universe through the atmospheric window (8–13 μm). It consumes no electricity and has great potential to be explored for cooling of buildings, vehicles, solar cells, and even thermal power plants. Radiative sky cooling has been explored in the past few decades but limited to nighttime use only. Very recently, owing to the progress in nanophotonics and metamaterials, daytime radiative sky cooling to achieve subambient temperatures under direct sunlight has been experimentally demonstrated. More excitingly, the manufacturing of the daytime radiative sky cooling material by the roll-to-roll process makes large-scale deployment of the technology possible. This work reviews the fundamental principles of radiative sky cooling as well as the recent advances, from both materials and systems point of view. Potential applications in different scenarios are reviewed with special attention to technology viability and benefits. As the energy situation and environmental issues become more and more severe in the 21st century, radiative sky cooling can be explored for energy saving in buildings and vehicles, mitigating the urban heat island effect, resolving water and environmental issues, achieving more efficient power generation, and even fighting against the global warming problem.

The world-wide energy demand for cooling and air conditioning is increasing dramatically in the 21st century due to the global warming effect, population growth, industrial development, and the improvement of living standards in the emerging and developing economies, especially those in hot climates.1 The current predominant vapor compression-based cooling technologies are facing problems such as massive energy consumption and the global warming effect due to the refrigerants used, e.g., hydrofluorocarbons (HFCs), which in turn causes environmental concerns.2,3 In addition, with cooling and refrigeration overwhelmingly produced through thermodynamic cycles by consuming nonrenewable fossil resources, cooling actually makes the earth hotter.4 The growing awareness of the energy situation and environmental concerns has led to much more interest in improving the efficiency of existing cooling systems and pursuing new alternative cooling technologies. It is well-known that all forms of matters emit radiation, and according to the Stefan-Boltzmann law, the higher the temperature of an object, the stronger its emissive power. Radiative cooling is a ubiquitous process by which a surface loses heat through thermal radiation. Considering that the earth has a surface temperature at around 300 K, while the cosmic microwave background of the universe has a thermal blackbody radiation spectrum at a temperature of 2.7 K,5 the large temperature difference between the earth and the universe can potentially be utilized to cool the earth surface by emitting thermal infrared radiation to the universe through the atmosphere. Unlike conventional cooling technologies that dump waste heat into the surroundings, including the local atmosphere and water bodies on the earth, radiative sky cooling sends excessive heat to the outer space without any energy consumption, making it truly appealing.6–12 

The significance of radiative sky cooling has been long recognized due to its “free” nature and the applications of radiative sky cooling can be traced back several centuries ago.13,14 However, it was not until the 1960s that this phenomenon was investigated systematically. Research studies on radiative sky cooling can generally be grouped into two categories, promoting the understanding of this phenomenon and applying the technique for real-world applications.15 The first category investigates the fundamental principles, e.g., emissivity properties of radiative cooling surfaces,16,17 spectral radiance of the atmosphere, and the dependence of radiative cooling on the wavelength,18–20 incident angle,21–23 and geographic locations.24–27 The second category focuses on finding new materials with desirable radiative properties and exploring radiative sky cooling for different application scenarios,28–33 e.g., cooling of residential and commercial buildings,34–39 cooling of solar cells,35,40–43 dew harvesting,44–48 outdoor personal thermal management,49 and supplemental cooling for condensers of air conditioners and power plants.50–53 The purpose of this work is to present a comprehensive review on radiative sky cooling, from the fundamental principles to materials development and applications in the terrestrial environment. Recent progresses on daytime radiative sky cooling are highlighted. This work will show that radiative sky cooling provides a potential pathway to approach energy saving, water saving, and more efficient power generation in various real-world applications.

Earlier research studies on radiative sky cooling have been limited to nighttime (i.e., nocturnal) use because even a few percent of solar absorption (solar intensity around 1000 W/m2) during the day could balance out the outgoing radiative cooling power (100–150 W/m2 depending on the surface temperature). There has been very active research on nighttime radiative sky cooling materials and devices in the 1970s and 1980s. Cooling of a surface at night to 10–15 °C below ambient in delicate devices has been experimentally demonstrated.54,55 However, the intrinsically low-energy-density of radiative sky cooling prohibits the widespread adoption of this technology. In general, radiative sky cooling can provide on average 40–80 W/m2 of net cooling power at night with clear sky conditions,36,37,56 which suggests that relatively large surface area is needed in order to deliver a meaningful cooling power, e.g., a few kilowatts (kWs) or larger. The need of a large surface area for radiative sky cooling in building applications is associated with the inevitably high installation and maintenance cost. Smith and Granqvist6 mentioned that the radiative sky cooling field has fallen short of its potential by a wide margin since few scientists are active in this technical area and the diverse technological scope is not well understood. To date, the most appealing radiative sky cooling application is high-albedo (i.e., high solar reflection) paints for cool roofs, which requires zero energy input and very limited maintenance but the energy savings can be very attractive.57 

For several decades, daytime (i.e., diurnal) radiative sky cooling remains to be a grand challenge due to the stringent requirements on materials, which should possess extremely low absorptivity in the high intensive solar spectrum (0.3–2.5 μm), while having high emissivity in other wavelengths, preferably within the atmospheric window (8–13 μm), which is the spectrum range that majority of thermal radiation from the earth's surface passing through the atmosphere to the universe. The recent demonstration of daytime radiative sky cooling has attracted great interest worldwide and revived the radiative sky cooling research field.31,32,58 Achieving subambient temperatures under the sun during the day is not only a significant scientific achievement in materials science, but also a meaningful milestone in engineering applications. The cooling energy harnessed during the day matches well with the cooling load profile because most end-users, e.g., buildings, have their peak cooling load taking place during the day. In addition, with the daytime radiative sky cooling demonstrated, a radiative sky cooling system can be built to run 24 h continuously, which renders even larger energy savings from the system.

In this review article, we begin our discussion by introducing the fundamental physics of radiative sky cooling in Sec. II. In Sec. III, radiative sky cooling materials for both nighttime and daytime are reviewed. Detailed technical constraints and prospects for each research area are discussed. Section IV reviews some exploration for real-world applications, along with the associated challenges.

Solar irradiation and emitted infrared radiation from the earth's surface comprise what is commonly termed environmental radiation.59 The interaction of environmental radiation with the earth's atmosphere maintains the energy balance and determines the temperature of the earth's surface [see Fig. 1(a)]. The incoming short wavelength (0.3–2.5 μm) solar irradiation can be reflected, absorbed, or scattered by air molecules and clouds in the atmosphere. In addition to solar irradiation, the atmosphere is also irradiated by the earth's surface, which is the terrestrial emission. Assuming a temperature of 300 K at the earth's surface, this upward terrestrial radiation is concentrated at wavelengths ranges from 2.5 to 50 μm according to the Planck's law. Similar to the downward solar irradiation, the upward radiation also experiences absorption and scattering when propagating through the atmosphere. Figure 1(b) is a simplified description of the energy flows at the earth's surface. For a surface on the ground that is exposed to the atmosphere, the net radiation effect is a combination of absorption of solar irradiation (during the day) and terrestrial thermal radiation. The net radiative cooling power (Pnetrad) can be expressed as

(1)

where Prad is the thermal radiation power from the surface. Patm is the absorbed atmospheric radiation power on the surface. Psolar is the absorbed solar irradiation power on the surface. These three parameters shown in the right-hand-side of Eq. (1) will be discussed in detail in Secs. II BII D. Section II A gives an introduction to atmospheric radiation in the wavelength range between 0.3 and 50 μm. Section II B focuses on the short wavelength range between 0.3 μm and 2.5 μm, and how the atmosphere affects the absorption of solar irradiation (Psolar) at the ground level. Section II C discusses the absorbed atmospheric radiation in the wavelength range between 3 μm and 50 μm (Patm). Section II D introduces thermal radiation from a solid surface (Prad).

FIG. 1.

Fundamentals of radiative sky cooling. (a) On a global scale, the earth gains energy from the sun and radiates the same amount of thermal energy to the universe to maintain its energy balance. (b) Heat transfer processes on a radiative cooling surface. Psolar (orange arrow) is the absorbed solar irradiation power on the surface. Patm (light blue arrow) is the absorbed atmospheric radiation power on the surface. Prad (dark red arrow) is the thermal radiation power from the surface. Pnon-radiative (green circle arrow) denotes the nonradiative heat transfer (convection and conduction) between the radiative cooling surface and the ambient.

FIG. 1.

Fundamentals of radiative sky cooling. (a) On a global scale, the earth gains energy from the sun and radiates the same amount of thermal energy to the universe to maintain its energy balance. (b) Heat transfer processes on a radiative cooling surface. Psolar (orange arrow) is the absorbed solar irradiation power on the surface. Patm (light blue arrow) is the absorbed atmospheric radiation power on the surface. Prad (dark red arrow) is the thermal radiation power from the surface. Pnon-radiative (green circle arrow) denotes the nonradiative heat transfer (convection and conduction) between the radiative cooling surface and the ambient.

Close modal

It is important to note that the radiative heat exchange in Fig. 1(b) and Eq. (1) is only between the radiative cooling surface, the atmosphere, and the sun, while neglecting the effect of the surrounding buildings, trees, and the ground. This usually applies to those surfaces that are placed horizontally or with a small tilt angle (e.g., a low-slope roof). For radiative sky cooling applications, the assumption of a horizontal surface toward the atmosphere is generally acceptable. However, in cases where the surface has a large tilt angle or vertically mounted, radiative heat exchange with the surrounding objects should be taken into consideration.37,60

When considering the nonradiative heat transfer processes from the surrounding, such as convection and conduction, the net cooling power (Pnet) can be expressed as

(2)

where Pnonradiative denotes the nonradiative heat transfer, i.e., conduction and convection, between the radiative cooling surface and the ambient, which will be discussed in detail in Sec. II E 5.

The study of atmospheric radiation (i.e., sky radiation) can be dated back to the early 19th century and is still very active today due to its importance in meteorology,61 remote sensing,62 and climate studies,63 to name a few. Pioneering studies on atmospheric radiation have been performed by researchers including Angstrom,64 Brunt,65 Bennett,66 Goody,67 Bell et al.,20 Bliss,68 Swinbank,69 Idso and Jackson,70 Brutsaert,71 and Martin and Berdahl.23 

Under clear sky conditions, the atmospheric radiation is a combined effect from different radiative sources, primarily different gas molecules and aerosol particles. Though the major components of the atmosphere are nitrogen (N2) and oxygen (O2), they have much less absorptivity/emissivity in the infrared spectra comparing to the major contributors of atmospheric absorption such as water vapor (H2O), carbon dioxide (CO2), and ozone (O3). Figure 2(a) shows the spectral emissivity of the atmosphere for the wavelengths from 0.3 to 50 μm. It is observed that the atmosphere has a low emissivity in the visible spectra (0.38–0.74 μm), but has several emission bands in the infrared spectra. Figures 2(b), 2(c), and 2(d) show that O3 emission has a relatively narrow band at a wavelength around 9.6 μm, CO2 has a broad emission band centered at wavelengths around 15 μm, and water vapor has a strong split band centered at a wavelength of 6.3 μm and has substantial emission at a wavelength greater than 20 μm, respectively.74 It is estimated that the percentage contribution to the greenhouse effect from water vapor and clouds, CO2, CH4, and O3 are 36%–72%, 9%–26%, 4%–9%, and 3%–7%, respectively.75 The absorptivity of the atmosphere in the wavelength range between 0.3 μm and 2.5 μm affects the absorption of solar irradiation at the ground level, while the absorptivity of the atmosphere in the wavelength range between 2.5 μm and 50 μm affects ground thermal radiation passing through the atmosphere to the universe.

FIG. 2.

Emissivity of various gases in the atmosphere at wavelengths ranging from 0.3 μm to 50 μm. (a) Total atmospheric emissivity. (b) Ozone emission has a relatively narrow band at wavelengths around 9.6 μm. (c) Carbon dioxide has three large emissivity bands in the infrared region at about 2.7, 4.3, and 15 μm. (d) Water vapor has several emissivity bands between 0.7 μm and 8 μm wavelengths and for wavelengths larger than 13 μm. Adapted with permission from Cold Facts on Global Warming, http://www.randombio.com/co2.html for absorption of ultraviolet, visible, and infrared radiation by various gases in the atmoshpere; accessed 7 August 2018. Copyright 2018 American Institute of Physics.72 From Physics of Climate. Copyright 1992 American Institute of Physics. Adapted with permission from AIP-Press.73 

FIG. 2.

Emissivity of various gases in the atmosphere at wavelengths ranging from 0.3 μm to 50 μm. (a) Total atmospheric emissivity. (b) Ozone emission has a relatively narrow band at wavelengths around 9.6 μm. (c) Carbon dioxide has three large emissivity bands in the infrared region at about 2.7, 4.3, and 15 μm. (d) Water vapor has several emissivity bands between 0.7 μm and 8 μm wavelengths and for wavelengths larger than 13 μm. Adapted with permission from Cold Facts on Global Warming, http://www.randombio.com/co2.html for absorption of ultraviolet, visible, and infrared radiation by various gases in the atmoshpere; accessed 7 August 2018. Copyright 2018 American Institute of Physics.72 From Physics of Climate. Copyright 1992 American Institute of Physics. Adapted with permission from AIP-Press.73 

Close modal

The sun emits approximately as a blackbody with an effective temperature of 5800 K. After sunlight rays travel through the atmosphere, sunlight is attenuated due to the scattering and absorption by atmospheric constituents, aerosols, and clouds. For a specific wavelength, the higher the absorptivity/emissivity of the atmosphere [see Fig. 2(a)], the greater the attenuation of the sunlight. In addition, the more atmosphere through which the sunlight passes, the greater the attenuation. The air mass (AM) coefficient76 can be used to characterize the solar spectrum after the sunlight has traveled through the atmosphere, which is defined as the direct optical path length through the atmosphere and expressed as a ratio relative to the path length vertically upward, i.e., the zenith. Figure 3 illustrates the spectral distribution of AM1.5 solar irradiation, which corresponds to a solar zenith angle of 48.2°. It can be observed that the solar irradiation is mostly concentrated in the short wavelength region (0.3–2.5 μm) with the peak occurring at approximately 0.5 μm in the visible range.

FIG. 3.

Absorption of solar irradiance on a horizontal radiative cooling surface. (a) Air mass (AM) 1.5 direct solar spectrum (highlighted in orange) and77 blackbody spectrum at a temperature of 5800 K (solid black curve). After sunlight travels through the atmosphere, sunlight is attenuated due to the scattering and absorption by atmospheric constituents. (b) Illustration of solar absorption on a horizontal surface.

FIG. 3.

Absorption of solar irradiance on a horizontal radiative cooling surface. (a) Air mass (AM) 1.5 direct solar spectrum (highlighted in orange) and77 blackbody spectrum at a temperature of 5800 K (solid black curve). After sunlight travels through the atmosphere, sunlight is attenuated due to the scattering and absorption by atmospheric constituents. (b) Illustration of solar absorption on a horizontal surface.

Close modal

The absorbed solar power is given by

(3)

where θ is the angle between the incident direction of solar irradiation and the normal direction of the surface, εsθ,λ is the emissivity of the surface as a function of direction and wavelength, and Isolarλ is the direct spectral solar irradiance.

For a long time, absorption of solar irradiation was not a concern in radiative sky cooling applications because they have been focused on nighttime use only. However, for application of radiative sky cooling during daytime under direct sunlight, solar absorption on a surface must be considered. Fortunately, the solar irradiation has been extremely well studied. Figure 3(a) shows that the wavelength of solar irradiation has a very small overlap with blackbody radiation at 300 K (black curve in Fig. 4). It is thus possible to develop a spectrally-selective surface that possess low emissivity close to zero (highly reflective) in the solar spectrum, while maintaining high emissivity in the infrared region for daytime radiative sky cooling.

FIG. 4.

The spectrum of a blackbody surface with a temperature of 300 K (solid black curve) and the atmospheric transmittance in the mid- and far-infrared regions (highlighted in blue). The wavelength range between 8 μm and 13 μm is usually termed as the atmospheric window due to the high transmittance. Coincidentally, the spectrum of atmospheric window matches well with the blackbody radiation curve from an object that has a temperature at around 300 K. For a clear sky with a low humidity level, the net radiative cooling power given by Eq. (1) could be as high as 140 W/m2 at ambient temperature (300 K).32 In addition, there exists a secondary atmospheric transmission window between 16 μm and 25 μm.78 The presence of the secondary transmission window has the potential to generate an extra of 10–20 W/m2 radiative cooling power.79 

FIG. 4.

The spectrum of a blackbody surface with a temperature of 300 K (solid black curve) and the atmospheric transmittance in the mid- and far-infrared regions (highlighted in blue). The wavelength range between 8 μm and 13 μm is usually termed as the atmospheric window due to the high transmittance. Coincidentally, the spectrum of atmospheric window matches well with the blackbody radiation curve from an object that has a temperature at around 300 K. For a clear sky with a low humidity level, the net radiative cooling power given by Eq. (1) could be as high as 140 W/m2 at ambient temperature (300 K).32 In addition, there exists a secondary atmospheric transmission window between 16 μm and 25 μm.78 The presence of the secondary transmission window has the potential to generate an extra of 10–20 W/m2 radiative cooling power.79 

Close modal

For radiative sky cooling applications, one usually has particular interest in the atmospheric emissivity (or transmittance) in the mid- and far-infrared regions (3–50 μm wavelength) because this is the emission spectrum of a blackbody at around 300 K (close to the ground temperature), as shown in Fig. 4. Also shown in Fig. 4 is the clear sky atmospheric transmittance in the mid- and far-infrared regions. The combined effects from all atmospheric components result in a minimum atmospheric radiation wavelength interval between 8 μm and 13 μm, i.e., the “atmospheric window.” Most of the terrestrial thermal radiation that can propagate through the atmosphere is in this wavelength range, where the atmosphere is highly transparent. Interestingly, the spectrum of the atmospheric window coincides with the wavelength range of thermal radiation from terrestrial objects. This indeed tells well how the earth keeps its surface temperature relatively stable throughout the year. It absorbs the sunlight but releases the thermal energy to the universe mostly by thermal radiation through the atmospheric window.

The absorption of atmospheric radiation on a horizontal surface facing toward the sky comes primarily from those wavelengths outside of the atmospheric window. According to the Kirchhoff's law of thermal radiation,59 under thermal equilibrium the emissivity value at a given wavelength and in a given direction equals to the absorptivity value in the same wavelength and direction, that is εsλ,θ=αsλ,θ. The total absorbed atmospheric radiation on a ground-level radiative cooling surface can be expressed as

(4)

where dΩ=0π/2dθsinθ02πdϕ is the angular integral over a hemisphere, εatmΩ,λ is the emissivity of atmosphere as a function of direction and wavelength, which is the ratio of the actual atmospheric radiation to the corresponding blackbody radiance at a certain angle and wavelength. Ibbλ,Tamb is the blackbody spectral radiance for a temperature Tamb, which can be further expressed as 2hc2λ5exphc/λkTamb1 according to Planck's law, where h=6.626×1034Js and k=1.381×1023J/K are the universal Planck and Boltzmann constants, respectively, c=2.998×108 m/s is the speed of light in vacuum, and λ is the wavelength. Tamb is the absolute air temperature near the ground. It is important to note that most of the atmospheric thermal radiation comes from the lowest few hundred meters of the atmosphere for those wavelengths where water vapor and carbon dioxide are strongly absorbing.54 Therefore, Tamb is usually used to characterize atmospheric radiation, which is a common practice in estimating downward (toward the ground) atmospheric radiation power.31,52

1. Characteristics of atmospheric emissivity

It is rather clear that atmospheric emissivity has spectral dependence as depicted in Fig. 2. Atmospheric emissivity is also dependent on several other factors such as the zenith angle,80,81 air humidity,82 cloud cover,83 aerosol,84 latitude,85 day-night effect,86,87 and seasonal effect.61,88 Here, we focus on introducing the dependence of atmospheric emissivity on the zenith angle and air humidity, which are the two most important factors. Figure 5(a) shows the atmospheric emissivity at three different zenith angles (0°, 60°, and 75°) in the wavelength range between 5 μm and 25 μm. Essentially, higher emissivity can be expected at larger zenith angles. A few earlier researchers have worked on expressing atmospheric emissivity as a function of zenith angle.22,23,74 For example, if atmospheric emissivity in the zenith angle εatm0,λ is known, atmospheric emissivity in other angles can be deduced by using a “box model,”54 which is given by εatmθ,λ=11εatm0,λ1/cosθ. Figure 5(b) shows the spectral emissivity as a function of the zenith angle and effective atmospheric emissivity (from 0.57 to 0.93). The emissivity values are not sensitive to zenith angle when zenith angles are small, and the emissivity values become unity (close to 1) when zenith angles are large.

FIG. 5.

Angular dependence of atmospheric emissivity. (a) The atmosphere gradually becomes more emissive as the zenith angle increases from 0° to 75°.54 Adapted with permission from J. Appl. Phys. 52, 4205 (1981). Copyright 1981 American Institute of Physics. (b) The spectral emissivity as a function of the zenith angle and total emissivity (from 0.57 to 0.93).81 Adapted with permission from X. Berger and J. Bathiebo, Renewable Energy 28, 1925–1933 (2003). Copyright 2003 Elsevier. The insets show an infrared image and an optical image that depict large spectral emissivity change at large zenith angles, which provide some insights into the angular dependence of radiative effects. The images were taken at the rooftop of Engineering Center, University of Colorado Boulder, USA.

FIG. 5.

Angular dependence of atmospheric emissivity. (a) The atmosphere gradually becomes more emissive as the zenith angle increases from 0° to 75°.54 Adapted with permission from J. Appl. Phys. 52, 4205 (1981). Copyright 1981 American Institute of Physics. (b) The spectral emissivity as a function of the zenith angle and total emissivity (from 0.57 to 0.93).81 Adapted with permission from X. Berger and J. Bathiebo, Renewable Energy 28, 1925–1933 (2003). Copyright 2003 Elsevier. The insets show an infrared image and an optical image that depict large spectral emissivity change at large zenith angles, which provide some insights into the angular dependence of radiative effects. The images were taken at the rooftop of Engineering Center, University of Colorado Boulder, USA.

Close modal

The level of water content in the atmosphere also greatly influences atmospheric emissivity in the atmospheric window (8–13 μm). Grant90 presented a critical review on the water vapor absorption coefficient in the 8–13 μm spectral region. Usually, the level of water content in the atmosphere is characterized as “precipitable water (PW, unit: millimeter or centimeter),” which is defined as the depth of water in a column assuming that all the water in that column is precipitated as rain.91Figure 6(a) shows the atmospheric emissivity under three different precipitable water levels, 1.5 cm, 3 cm, and 6 cm, respectively. Figure 6(b) gives the relationship between precipitable water and relative humidity at various ambient temperatures. It can be found that atmospheric emissivity increases with the precipitable water, and the presence of water vapor reduces the radiative cooling power.92 By using a similar setup as Raman et al.,31 who has obtained 40.1 W/m2 daytime radiative cooling power at Stanford, California, Tso et al.93 conducted a field test in Hong Kong and showed that radiative sky cooling does not perform as well under a humid and cloudy climate. The researchers found that the maximum nighttime cooling power was only 38 W/m2 with a clear sky and did not observe any daytime cooling effect. Suichi et al.94 conducted tests in Okayama, Japan (warm and humid) by using alternating layers of SiO2 and poly(methyl methacrylate) on an aluminum mirror. It was shown that the device is 2.8 °C higher than the ambient temperature of 35 °C during the day. With water vapor being the most important source of infrared absorption, the atmospheric emissivity depends strongly on the geographic location and climate conditions.

FIG. 6.

The dependence of atmospheric emissivity on humidity. (a) The atmosphere becomes more emissive as the precipitable water vapor content increases from 1.5 cm to 6 cm.54 The red solid curve corresponds to the red solid curve in Fig. 5(a). Adapted with permission from J. Appl. Phys. 52, 4205 (1981). Copyright 1981 American Institute of Physics. (b) Precipitable water as a function of relative humidity at various ambient temperatures. Results calculated using equations from Ref. 89.

FIG. 6.

The dependence of atmospheric emissivity on humidity. (a) The atmosphere becomes more emissive as the precipitable water vapor content increases from 1.5 cm to 6 cm.54 The red solid curve corresponds to the red solid curve in Fig. 5(a). Adapted with permission from J. Appl. Phys. 52, 4205 (1981). Copyright 1981 American Institute of Physics. (b) Precipitable water as a function of relative humidity at various ambient temperatures. Results calculated using equations from Ref. 89.

Close modal

2. Prediction of atmospheric emissivity

Spectral atmospheric emissivity [εatmλ,θ] is very complicated due to the different constituents of the atmosphere, along with the varying weather and climate conditions. Atmospheric emissivity for a particular location can be obtained either from direct radiance measurements by using a spectral radiometer,18,22,95,96 or more commonly, can be obtained by modeling methods based on ground-level meteorological data.19,27,97,98 Under clear sky conditions, the atmospheric spectral transmittance [1 − εatmλ,θ] can be calculated using several computational tools such as LOWTRAN (LOWTRAN 1–3,99–101 LOWTRAN 4,19,102 LOWTRAN 5,103,104 LOWTRAN 6,85,105 LOWTRAN 7,106,107), MODTRAN,108,109 and ATRAN.110,111 Note that both MODTRAN and ATRAN were developed and improved based on LOWTRAN.

In practice, a lumped “effective atmospheric emissivity”61,112 (ε¯atm, sometimes also named as sky emissivity,95,113,114 global sky emissivity,22 total atmosphere emissivity,81 total sky emissivity,105 integrated sky emissivity,115 or average atmosphere emissivity55) concept is usually defined based on the absolute air temperature near the ground, which assumes a continuous spectrum disregarding the actual spectral distribution of the outgoing and incoming radiation

(5)

when expressed in terms of a lumped effective atmospheric emissivity [defined in Eq. (5)], which is usually described by a linear or quadratic equation as a function of dew point temperature and/or ambient temperature,65,116 the effective atmospheric emissivity value can be used to describe the full spectrum, or a particular spectrum range of interest, for example, 8–13 μm (ε¯atm,813μm).117 A list of these correlations can be found in earlier review papers.34,118–122 However, when using these simple correlations, one has to carefully consider their validity conditions. For example, some of the equations are developed for specific conditions such as arid regions, at sea level, clear sky, or specific wavelength range, while others may consider one or more of the following conditions such as day-night effect,22 seasonal effect,88 and water vapor pressure,27,106 as summarized in Table I. In general, effective atmospheric emissivity values are sufficient to be used to calculate radiative cooling power in most scenarios. However, it is important to note that when using the effective atmospheric emissivity (ε¯atm) and effective emissivity of a radiative cooling surface (ε¯s) to calculate the absorbed atmospheric radiation on the surface, that is Patm=ε¯sε¯atmσTamb4, the ε¯atm could be ill-defined in some cases. The reason is that even if two surfaces have the same lumped emissivity but different spectral emissivity characteristics, they might not absorb the same amount of the atmospheric radiation. In these cases, spectral-dependent atmospheric emissivity is preferred.

TABLE I.

List of correlations that predict the effective atmospheric emissivity. Tdp is the dewpoint temperature (Kelvin); Tamb is the ambient temperature (Kelvin); p¯pw is the partial water pressure (millibar); Z is altitude (meter); Patm is atmospheric pressure (hectopascal); RH is relative humidity (percentage);clf is cloud fraction; and PW is precipitable water (millimeter).

CorrelationValidity conditionsYearSkyReference
εsky=a+bp¯pw0.5 Zenith angle dependent 1932 Clear Brunt65  
εsky=0.67p¯pw0.08 0.2<p¯pw<20mb 1972 Clear Staley and Jurica112  
εsky=0.1+3.53×108p¯pw2exp3000/Tamb 10.5–12.5 μ1981 Clear Idso117  
εsky=0.24+2.98×108p¯pw2exp3000/Tamb 8–14 μ1981 Clear Idso117  
εsky=0.601+5.95×108p¯pw2exp1500/Tamb For Antarctic and Arctic 1982 Clear Andreas and Ackley123  
εsky=5.7723+0.9555×0.6017z×Tamb1.893RH0.0665×104 Nighttime and altitude dependent 1982 Clear Centeno113  
εsky=0.741+0.0062Tdp273.15 Nighttime 1982 Clear Berdahl and Fromberg22  
εsky=0.727+0.0060Tdp273.15 Daytime 1982 Clear Berdahl and Fromberg22  
εsky=0.711+0.56Tdp/100+0.73Tdp/1002+0.013cos2πt/24+0.00012Patm1000 Nighttime and altitude dependent 1984 Clear Martin and Berdahl27  
εsky=0.770+0.0038Tdp Nighttime 1984 Clear Berger et al.124  
εsky=exp1.662.321.875Tamb/273.15+0.735PW/250.5  1998 Clear Dilley and O'Brien125  
εsky=0.72+0.009p¯pw2 p¯pw2 2001 Clear Niemelä et al.126  
εsky=0.72+0.076p¯pw2 p¯pw<2 2001 Clear Niemelä et al.126  
εsky=0.754+0.0044Tdp Nighttime, Negev highlands, Israel 2004 Clear Tang et al.114  
εsky=1.18p¯pw/Tamb0.143 For Andean Altiplano, Bolivia 2007 Clear Lhomme et al.127  
εsky=εsky,clear+N1D/Tamb4εsky,clear Altitude dependent, D is a constant related to altitude 1982 Cloudy Centeno113  
εsky=εsky,clear+C1εsky,clear where C is a constant related to cloud fraction 1984 Cloudy Martin and Berdahl27  
εsky=clf+1clfεsky,clear  1999 Cloudy Crawford and Duchon61  
εsky=1.180.34Psolar,realPsolar,cloud+1.37 For Andean Altiplano, Bolivia 2007 Cloudy Lhomme et al.127  
εsky=Cp¯pw/T1.670.83PshortwavePextra For Zongo, Bolivia 2010 Cloudy Sicart et al.128  
CorrelationValidity conditionsYearSkyReference
εsky=a+bp¯pw0.5 Zenith angle dependent 1932 Clear Brunt65  
εsky=0.67p¯pw0.08 0.2<p¯pw<20mb 1972 Clear Staley and Jurica112  
εsky=0.1+3.53×108p¯pw2exp3000/Tamb 10.5–12.5 μ1981 Clear Idso117  
εsky=0.24+2.98×108p¯pw2exp3000/Tamb 8–14 μ1981 Clear Idso117  
εsky=0.601+5.95×108p¯pw2exp1500/Tamb For Antarctic and Arctic 1982 Clear Andreas and Ackley123  
εsky=5.7723+0.9555×0.6017z×Tamb1.893RH0.0665×104 Nighttime and altitude dependent 1982 Clear Centeno113  
εsky=0.741+0.0062Tdp273.15 Nighttime 1982 Clear Berdahl and Fromberg22  
εsky=0.727+0.0060Tdp273.15 Daytime 1982 Clear Berdahl and Fromberg22  
εsky=0.711+0.56Tdp/100+0.73Tdp/1002+0.013cos2πt/24+0.00012Patm1000 Nighttime and altitude dependent 1984 Clear Martin and Berdahl27  
εsky=0.770+0.0038Tdp Nighttime 1984 Clear Berger et al.124  
εsky=exp1.662.321.875Tamb/273.15+0.735PW/250.5  1998 Clear Dilley and O'Brien125  
εsky=0.72+0.009p¯pw2 p¯pw2 2001 Clear Niemelä et al.126  
εsky=0.72+0.076p¯pw2 p¯pw<2 2001 Clear Niemelä et al.126  
εsky=0.754+0.0044Tdp Nighttime, Negev highlands, Israel 2004 Clear Tang et al.114  
εsky=1.18p¯pw/Tamb0.143 For Andean Altiplano, Bolivia 2007 Clear Lhomme et al.127  
εsky=εsky,clear+N1D/Tamb4εsky,clear Altitude dependent, D is a constant related to altitude 1982 Cloudy Centeno113  
εsky=εsky,clear+C1εsky,clear where C is a constant related to cloud fraction 1984 Cloudy Martin and Berdahl27  
εsky=clf+1clfεsky,clear  1999 Cloudy Crawford and Duchon61  
εsky=1.180.34Psolar,realPsolar,cloud+1.37 For Andean Altiplano, Bolivia 2007 Cloudy Lhomme et al.127  
εsky=Cp¯pw/T1.670.83PshortwavePextra For Zongo, Bolivia 2010 Cloudy Sicart et al.128  

The above discussion only considers clear sky conditions (though humidity level can be different). Yet, the sky could be covered by clouds for a significant amount of time during the day and throughout the year. Therefore, the effect of clouds on downward atmospheric radiation has also been investigated extensively.129–133 Unfortunately, the effect of cloud cover has very complicated impacts on the atmospheric radiation due to the different types of clouds (e.g., cirrus, stratus), different altitude of clouds, different thickness of clouds, different constitution of clouds, different droplet size distribution, different observation angle from the radiative cooling surface, and the constantly changing of cloud behaviors over time and space.129,134,135 Since the major composition of clouds is condensate water droplets and ice crystals that have an average emissivity of 0.92–0.98,21,136 the emissivity of thick clouds may be considered as 1 over the whole infrared spectrum, including the atmospheric window.135 

Experimental results show that the decrease in radiative cooling power could be proportional to the increase in cloud cover fraction, compared to clear sky, as shown in Fig. 7. However, the experiments shown in Fig. 7 only presents the effect of cloud cover fraction, while neglecting other factors such as cloud height, cloud thickness, and observation angle from the radiative cooling surface. This is the reason why net radiative cooling powers have a minor difference between cases “few” and “clear” in Fig. 7(a). A more quantitative study on the effect of cloud is still in the need. Usually, a sky clearness index (i.e., cloud cover fraction) can be obtained from meteorological data.137,138 Then, for a cloud cover fraction of f, the power of absorbed atmospheric radiation may be calculated as the sum of two terms respectively associated with clear and cloudy skies

(6)

where Tcloud is the cloud base temperature.

FIG. 7.

Measurement of net radiative cooling power on different days with different cloud cover conditions at Boulder, Colorado, USA. (a) The net radiative cooling power at different cloud cover conditions. The cloud conditions are estimated in terms of how many eighths of the sky are covered by the cloud, ranging from 0 (clear sky) to 8 (overcast), which is a common practice in meteorology.139 The human observations of the cloud amount are usually reported in five categories: clear (0/8), few (1/8–2/8), scattered (3/8–4/8), broken (5/8–7/8), and overcast (8/8). (b) Pictures show sky conditions during the test days.

FIG. 7.

Measurement of net radiative cooling power on different days with different cloud cover conditions at Boulder, Colorado, USA. (a) The net radiative cooling power at different cloud cover conditions. The cloud conditions are estimated in terms of how many eighths of the sky are covered by the cloud, ranging from 0 (clear sky) to 8 (overcast), which is a common practice in meteorology.139 The human observations of the cloud amount are usually reported in five categories: clear (0/8), few (1/8–2/8), scattered (3/8–4/8), broken (5/8–7/8), and overcast (8/8). (b) Pictures show sky conditions during the test days.

Close modal

It is important to note that even for a completely cloudy sky, the atmospheric window is completely opaque and radiative sky cooling is still possible because the cloud base temperature can be lower than the ground temperature. The cooling power in this case is primarily determined by the base height of the cloud. As the height of the cloud base increases, radiative cooling power also increases. If the cloud base height from the ground is known, the cloud base temperature can be estimated using the lapse rate of atmospheric temperature under humid conditions. The lapse rate varies from 2.5 °C/km to 7.5 °C/km depending on the ground temperature, season, and geographic location.140 If the cloud base height is low and the lapse rate at a specific location or during a specific time period is also low, the cloud temperature may be assumed equal to the ambient temperature. In this case, radiative cooling power would be close to zero.

The total radiative power per unit area (Prad) from a surface is the rate at which radiation is emitted at all possible wavelengths and in all possible directions, as given in the following equation:

(7)

where Ts is the absolute temperature of the radiative cooling surface, Ω is the solid angle between the direction of radiation and normal to the surface [Fig. 8(b)], and εsΩ,λ is the emissivity of the surface as a function of wavelength and direction.

FIG. 8.

Fundamentals of thermal radiation from a surface. (a) Spectral distribution of blackbody and real surface emission. (b) Directional distribution of blackbody and real surface emission. (c) An example showing reflectance [RΩ,λ=1εsΩ,λ, transmission = 0] of doped polyethylene foil on aluminum as a function of wavelength and angle of incidence.142 Reproduced with permission from A. R. Gentle and G. B. Smith, Nano Lett. 10, 373–379 (2010). Copyright 2010 American Chemical Society.

FIG. 8.

Fundamentals of thermal radiation from a surface. (a) Spectral distribution of blackbody and real surface emission. (b) Directional distribution of blackbody and real surface emission. (c) An example showing reflectance [RΩ,λ=1εsΩ,λ, transmission = 0] of doped polyethylene foil on aluminum as a function of wavelength and angle of incidence.142 Reproduced with permission from A. R. Gentle and G. B. Smith, Nano Lett. 10, 373–379 (2010). Copyright 2010 American Chemical Society.

Close modal

The most important parameter in Eq. (7) is the emissivity of the surface εsΩ,λ, which is usually assumed to be independent of temperature because the variation of surface emissivity in the temperature range of interest for most radiative sky cooling applications is negligible. Thus, it is reasonable to assume that the surface emissivity εsΩ,λ has angular and spectral dependence only, as shown in Figs. 8(a) and 8(b). The emissivity for a specific material may have different values at a given wavelength or in a given direction, or in integrated averages over a particular wavelength band and in certain direction. For example, in radiative sky cooling applications, due to the importance of the atmospheric window, particular interest has been paid to the integrated average emissivity in the range 8 < λ < 13 μm and in the zenith direction. Figure 8(c) shows that a surface has low reflection (i.e., high emission) at a small angle of incidence, and as the angle of incidence approaches 90°, the reflection will become close to unity. For most materials, the emissivities usually remain relatively stable and high for an angle of incidence between 0° and 60°.31,141

Instead of spectral and wavelength dependence emissivity εsΩ,λ, the total hemispherical emissivity (ε¯s) is also used, which is defined as the total radiation energy emitted over all wavelengths and in all directions from a surface

(8)

where σ is the Stefan-Boltzmann constant, 5.67 × 10−8 W/(m2 K4).

Based on the total hemispherical emissivity, Eq. (7) can be rewritten as

(9)

1. Ideal emissivity curves

From what has been discussed above, one can obtain a few ideal surface emissivity curves for radiative sky cooling applications in different scenarios. Figure 9 shows the ideal emissivity spectrum of radiative sky cooling materials in four different application scenarios: daytime above-ambient, daytime subambient, nighttime above-ambient, and nighttime subambient. Almost all research efforts on radiative sky cooling materials have been devoted to design and synthesize materials with these spectral curves. The wavelength ranges that are of interest are 2.5–50 μm and 0.3–50 μm for nighttime and daytime applications, respectively. Subambient applications require emissivity close to unity in the atmospheric window (8–13 μm), and daytime applications require emissivity close to zero in the solar spectrum (0.3–2.5 μm).

FIG. 9.

Ideal emissivity spectrum of radiative sky cooling materials in four different application scenarios, daytime above-ambient (black solid line), daytime subambient (red dashed line), nighttime above-ambient (orange dashed line), and nighttime subambient (blue solid line), respectively.

FIG. 9.

Ideal emissivity spectrum of radiative sky cooling materials in four different application scenarios, daytime above-ambient (black solid line), daytime subambient (red dashed line), nighttime above-ambient (orange dashed line), and nighttime subambient (blue solid line), respectively.

Close modal

2. Thermal measurement methods and devices

While spectroscopic characterization of emissivity, absorptivity, reflectivity, and transmissivity of a material at a specific wavelength is a routine practice in materials sciences (see Sec. III), thermal measurement of a radiative cooling surface is not straightforward. The measurement can be conducted in different ways for different objectives. Figure 10 shows three types of measurement methods with the schematic of their corresponding devices. Figure 10(a) shows a radiative cooling surface that is directly exposed to ambient air. This type of measurement is usually for those applications that need to decrease the temperature of the underlying structure, such as a cool roof (see Sec. IV A).143 Typical measurement results from this method are given in Fig. 10(b), where a 2–5 °C subambient cooling is demonstrated with a glass-polymer hybrid metamaterial32 at noon (11 am–2 pm) under ∼800 W/m2 solar irradiance. Figure 10(c) shows a measurement device designed for obtaining large subambient temperatures, which is usually employed to demonstrate the effectiveness of a wavelength-selective surface. The convection shield is used to suppress nonradiative heat transfer (conduction and convection loss) between the radiative cooling surface (at subambient temperatures) and ambient. Usually, 10–15 °C subambient temperatures can be achieved at night with clear sky conditions and low humidity. By using a dedicated vacuum chamber that is shielded from direct sunlight, Chen et al.144 demonstrated an average temperature reduction of 37 °C from the ambient through a 24-h day-night cycle, as shown in Fig. 10(d). The devices shown in Figs. 10(a) and 10(c) focuses on demonstrating a temperature difference between the radiative cooling surface and ambient, but with a small thermal mass. In real-world applications, one might be interested in cooling an object with a decent amount of thermal mass. This type of measurement can be conducted using the device shown in Fig. 10(e). By placing the radiative cooling metamaterial surface32 onto a 9-mm-thick water layer, Zhao et al.52 demonstrated 10.6 °C subambient cooling of water under ∼700 W/m2 solar irradiance, as shown in Fig. 10(f).

FIG. 10.

Schematic drawings of three different thermal measurement devices for radiative sky cooling materials. (a) The radiative cooling surface (solid blue line) is placed on a substrate and directly exposed to the ambient. (b) Results of a typical cool roof measurement conducted at Laramie, Wyoming, USA. The experiment was performed on the roof of a single-space model room (2.44 m × 1.83 m × 2.44 m, L × W × H). (c) The radiative cooling surface is placed in a well-insulated enclosure (highlighted in gray). The insulation can be regular polystyrene boards or a dedicated vacuum chamber. A convection shield (solid yellow line) is employed at the top to suppress thermal losses due to convection and conduction heat transfer at subambient conditions. (d) Measurement results obtained using a dedicated vacuum chamber to minimize parasitic heat losses. An average temperature reduction of 37 °C from the ambient air temperature is achieved at Stanford, California, USA.144 (e) Instead of demonstrating a temperature difference between the radiative cooling surface and ambient (TambTs), the radiative cooling surface is placed on an object (e.g., water) that has a large thermal mass to generate useful cooling energy. (f) Measurement results obtained from using a radiative cooling module (0.58 m × 0.58 m surface area) that contains a 9-mm-thick stationary water layer. The measurement was conducted at Boulder, Colorado, USA.

FIG. 10.

Schematic drawings of three different thermal measurement devices for radiative sky cooling materials. (a) The radiative cooling surface (solid blue line) is placed on a substrate and directly exposed to the ambient. (b) Results of a typical cool roof measurement conducted at Laramie, Wyoming, USA. The experiment was performed on the roof of a single-space model room (2.44 m × 1.83 m × 2.44 m, L × W × H). (c) The radiative cooling surface is placed in a well-insulated enclosure (highlighted in gray). The insulation can be regular polystyrene boards or a dedicated vacuum chamber. A convection shield (solid yellow line) is employed at the top to suppress thermal losses due to convection and conduction heat transfer at subambient conditions. (d) Measurement results obtained using a dedicated vacuum chamber to minimize parasitic heat losses. An average temperature reduction of 37 °C from the ambient air temperature is achieved at Stanford, California, USA.144 (e) Instead of demonstrating a temperature difference between the radiative cooling surface and ambient (TambTs), the radiative cooling surface is placed on an object (e.g., water) that has a large thermal mass to generate useful cooling energy. (f) Measurement results obtained from using a radiative cooling module (0.58 m × 0.58 m surface area) that contains a 9-mm-thick stationary water layer. The measurement was conducted at Boulder, Colorado, USA.

Close modal

3. Broadband and selective radiative sky cooling surfaces

To achieve a large emissive power, a surface should have emissivity close to unity for all wavelengths, i.e., a broadband emitter, and in all directions. However, for a broadband emitter, according to the Kirchhoff's law of thermal radiation,59 the spectral absorption equals to the spectral emission at thermodynamic equilibrium. Therefore, for subambient radiative cooling, the radiative cooling surface would also absorb radiation from the surroundings (e.g., ground, buildings, and trees) which are at temperatures much higher than the temperature of atmosphere or the universe. To reduce the incoming radiative flux, radiative cooling surfaces need to be designed to have spectral dependent emissivity (i.e., selective surface), that is, the emissivity at some specific wavelengths is higher than other regions, for example, within the atmospheric window (i.e., 8–13 μm). The effectiveness of a wavelength-selective surface can be defined as54 

(10)

where ε¯s,813μm and ε¯s,0 denote the integral of the emissivity over the atmospheric window and the whole spectrum, respectively.

Since the incoming radiative flux is minimized, a selective surface could achieve a much lower temperature compared to a broadband surface.55Figure 11 shows an example of performance comparison between the broadband and wavelength-selective surfaces. Figure 11(a) shows an ideal broadband surface (emissivity = 1 in the wavelength range between 3 μm and 25 μm), ideal wavelength-selective surface (emissivity =1 in the wavelength range between 8 μm and 13 μm, and emissivity = 0 elsewhere), real broadband surface, and real wavelength-selective surface, respectively. The real broadband surface is a 0.5-mm-thick polycarbonate sheet, and the real wavelength-selective surface is a 30–μm-thick polyvinylidene fluoride (PVDF) thin film. Figure 11(b) shows the theoretically calculated net cooling power for the 4 surfaces. Clearly, for both ideal and real surfaces, the broadband surfaces have a larger net cooling power when TsTamb>0 or TambTs is small, while selective surfaces can achieve a larger subambient temperature drop. The results of theoretical calculation are confirmed by a comparison test between the PVDF thin film and the polycarbonate sheet in Boulder, Colorado, USA, as shown in Fig. 11(c). The two surfaces are controlled to have the same initial temperature. At the beginning, the rate of temperature decrease is faster for the broadband surface, while at the “steady-state,” temperature differences TambTs are 16.0 °C and 14.7 °C for the selective surface and the broadband surface, respectively. The test results clearly confirm that the broadband surface has larger cooling power at temperatures close to the ambient and the selective surface can achieve lower temperature. Figure 11(d) gives the precipitable water and wind speed conditions during the test.

FIG. 11.

Broadband blackbody and wavelength-selective radiative cooling surfaces. (a) Emissivity spectrum for ideal broadband, ideal wavelength-selective, real broadband, and real wavelength-selective surfaces. The real broadband surface is a 0.5-mm-thick polycarbonate sheet, and the real wavelength-selective surface is a 30-μm-thick PVDF thin film. (b) Theoretical analysis of the nighttime net cooling power as a function of the temperature difference between radiative cooling surface and ambient (TsTamb) for the four types of surfaces. (c) Comparison test of the real broadband and real wavelength-selective surfaces. Temperatures of the two surfaces are plotted as a function of time during September 13–14, 2018, at Boulder, Colorado, USA. The initial temperatures of the two surfaces are controlled to be identical. It is clear that the broadband surface has a larger cooling power when TsTamb is small, but the wavelength-selective surface can reach a lower temperature. The temperature curves are measured using a thermal testing device with the schematic shown in Fig. 10(c). (d) Precipitable water and wind speed conditions for the comparison test.

FIG. 11.

Broadband blackbody and wavelength-selective radiative cooling surfaces. (a) Emissivity spectrum for ideal broadband, ideal wavelength-selective, real broadband, and real wavelength-selective surfaces. The real broadband surface is a 0.5-mm-thick polycarbonate sheet, and the real wavelength-selective surface is a 30-μm-thick PVDF thin film. (b) Theoretical analysis of the nighttime net cooling power as a function of the temperature difference between radiative cooling surface and ambient (TsTamb) for the four types of surfaces. (c) Comparison test of the real broadband and real wavelength-selective surfaces. Temperatures of the two surfaces are plotted as a function of time during September 13–14, 2018, at Boulder, Colorado, USA. The initial temperatures of the two surfaces are controlled to be identical. It is clear that the broadband surface has a larger cooling power when TsTamb is small, but the wavelength-selective surface can reach a lower temperature. The temperature curves are measured using a thermal testing device with the schematic shown in Fig. 10(c). (d) Precipitable water and wind speed conditions for the comparison test.

Close modal

We note here both broadband and selective surfaces are important depending on the applications whether a larger net cooling power or a lower temperature is more favorable. In general, a broadband surface has larger cooling power at temperatures higher than (above-ambient) or close to ambient temperature, e.g., cooling of solar cells, while a selective surface could achieve larger subambient temperature reduction.145 Granqvist146 summarized materials for nighttime radiative sky cooling and concluded that for most of the time, a broadband surface works efficiently due to their larger cooling power.

4. Effect of surface temperature on the net radiative cooling power

It is important to understand how the net radiative cooling power changes with the surface temperature. Figure 12 shows the theoretical and experimental results of the net radiative cooling power [given in Eq. (1)] as a function of surface temperature.52 Here, all tests were conducted under the condition that the radiative cooling surface temperature is controlled to be close to the ambient temperature, so the nonradiative heat transfer can be ignored. Essentially, a higher surface temperature gives a higher net radiative cooling power for both day- and night time. This is consistent with the Stefan-Boltzmann's law, the higher the temperature, the larger the emissive power of a surface. At night, when the radiative cooling surface temperature is increased from −1 to 20.5 °C, the net radiative cooling power increases from 61.2 to 101.4 W/m2. Therefore, over 100 W/m2, the net radiative cooling power can be easily expected with a surface temperature higher than 20 °C. Daytime experiments were performed at around noon with a direct solar irradiance on the radiative cooling surface between 680 and 730 W/m2. When the surface temperature is increased from 10.9 to 31.9 °C, the net radiative cooling power increased from 39.7 to 79.0 W/m2. For the modeling part, the absorbed solar irradiation power on the surface (Psolar), the absorbed atmospheric radiation power on the surface (Patm), and the thermal radiation power from the surface (Prad), are obtained from Eqs. (3), (4), and (7), respectively. These three parameters are then substituted into Eq. (1) to obtain the net radiative cooling power. The calculated net radiative cooling power in Fig. 12 agrees well with the measurement results for both daytime and nighttime. Therefore, correlations between the net radiative cooling power and surface temperature are developed for daytime and nighttime, respectively: Pnetrad=0.0079×Ts2+1.27×Ts+31.6 and Pnetrad=0.0079×Ts2+1.27×Ts+69.9 (valid for surface temperatures between −10 °C and 40 °C).

FIG. 12.

Net radiative cooling power as a function of surface temperature. All tests were conducted at Boulder, Colorado, under the condition that the surface temperature is close to the ambient temperature to eliminate the nonradiative heat transfer between radiative cooling surface and the ambient. Comparison of experimentally measured net radiative cooling power (dots) to the modeling results (curves) under clear sky conditions, with measured 7.4–10.7 mm precipitable water at night and 12.5–17.0 mm precipitable water during the day. Here, the direct solar irradiance on the surface during the daytime is assumed to be 700 W/m2. Reproduced with permission from Zhao et al., Joule 3, 111–123 (2019). Copyright 2018 Elsevier Inc.

FIG. 12.

Net radiative cooling power as a function of surface temperature. All tests were conducted at Boulder, Colorado, under the condition that the surface temperature is close to the ambient temperature to eliminate the nonradiative heat transfer between radiative cooling surface and the ambient. Comparison of experimentally measured net radiative cooling power (dots) to the modeling results (curves) under clear sky conditions, with measured 7.4–10.7 mm precipitable water at night and 12.5–17.0 mm precipitable water during the day. Here, the direct solar irradiance on the surface during the daytime is assumed to be 700 W/m2. Reproduced with permission from Zhao et al., Joule 3, 111–123 (2019). Copyright 2018 Elsevier Inc.

Close modal

5. Nonradiative heat transfer

In addition to thermal radiation, the radiative cooling surfaces are also subject to other nonradiative heat transfer processes with the surrounding environment. The nonradiative loss can be lumped together as

(11)

where h is an overall heat transfer coefficient that accounts for convection and conduction heat transfer between the ambient and the radiative cooling surface.

To quantify nonradiative heat transfer, one needs to accurately determine the overall heat transfer coefficient h. Fortunately, wind flow over rectangular plate surfaces such as solar collectors or photovoltaic (PV) panels has been well-studied.147,148 However, it is not suggested to directly apply those h values for solar collectors or PV panels in radiative sky cooling applications, especially when the local wind speed is small (i.e., natural convection dominates). The reason is that solar collectors and PV panels usually have surface temperature much higher than ambient, while radiative cooling surfaces have temperatures close to or lower than the ambient. If without a convection shield, Zhao et al.52 showed that for a radiative sky cooling device that has a 0.58 m × 0.58 m surface area, the following equation can be used to quantify the nonradiative thermal loss due to wind:

(12)

where Vwind is the zero-incidence wind velocity. This equation is generally valid for local wind speed between 0 m/s and 8 m/s.

In case that there is a convection shield over the radiative cooling surface, the h should be case-specific due to the different convection shield structures and the air gap thickness between the convection shield and the surface. For a radiative sky cooling device (0.58 m × 0.58 m surface area) that has a polyethylene (PE) film 1.5-cm above the surface, Zhao et al.52 suggested the following overall heat transfer coefficient (valid for local wind speed between 0 m/s and 8 m/s):

(13)

The convection and conduction heat transfer can be either beneficial or detrimental to radiative sky cooling, depending on the operating temperature of a radiative cooling surface (Ts) compared to temperature of ambient (Tamb). In general, for above-ambient applications, e.g., cooling of solar cells149 and supplemental cooling of power plant's condensers,150 nonradiative heat transfer is beneficial to the cooling process. However, for subambient applications,6,51 nonradiative heat transfer must be suppressed.

At subambient operating conditions, the extent of subambient temperature reduction and the cooling power at a certain subambient temperature of the radiative cooling surface is limited by the heat exchange with surrounding warmer air. Figure 13(a) shows that different overall heat transfer coefficients have strong influence on the net cooling power and the lowest achievable temperature. For a typical radiative sky cooling device as shown in Fig. 9(c), the suppression of heat transfer from all sides and the bottom surface are necessary but reasonably straightforward by using highly insulating materials. However, to insulate the top surface that facing the sky where the infrared radiative flux comes out, the covering material (i.e., convection shield) must possess the following characteristics: (1) high transmittance close to unity in the whole infrared region, especially in the atmospheric window (8–13 μm), (2) high mechanical strength to survive strong winds and even hail, (3) low cost, and (4) long term durability. In some studies,151,152 the covering material also needs to have high reflectance in the solar spectrum in order to prevent heating caused by the sun during the day. In these cases, the reflection of solar radiation is done by the covering material, instead of the radiative cooling surface. However, no covering materials have yet been identified with all these desired features, while the best candidate till now is the polyethylene (PE) film that has a thickness usually less than 30 μm.92,115,142,153–155 The PE film is usually installed a few centimeters above the radiative cooling surface as a convection shield where the air is trapped between the PE and the surface to suppress convective and conductive thermal losses.

FIG. 13.

Convection shield to suppress parasitic thermal losses through convection and conduction to the ambient from a radiative sky cooling device. (a) Net cooling power plotted as a function of subambient temperatures under a variety of overall heat transfer coefficients. The 30-μm-thick PVDF film is employed as a radiative cooling surface. Deep subambient temperature ≥25 °C can be achieved if h is controlled to be ≤1 W/(m2 K). Results are obtained at ambient temperature 20 °C and precipitable water 10 mm. (b) Transmittance of the polyethylene (PE) film with different thicknesses in the wavelength range between 5 μm and 50 μm. Data adapted with permission from J. Appl. Phys. 52, 4205 (1981). Copyright 1981 American Institute of Physics, Nilsson et al., Sol. Energy Mater. 12, 327–333 (1985). Copyright 1985 Elsevier, Berdahl et al., Int. J. Heat Mass Transfer 26(6), 871–880 (1983). Copyright 1983 Pergamon Press Ltd. (c) Transmittance of the PE film, ZnS, and ZnSe in the infrared range. The solid red curve in (b) corresponds to the red solid curve in (c). Data adapted with permission from J. Appl. Phys. 52, 4205 (1981). Copyright 1981 American Institute of Physics, S. N. Bathgate and S. G. Bosi, Sol. Energy Mater. Sol. Cells 95, 2778–2785 (2011). Copyright 2011 Elsevier, Chen et al., Nat. Commun. 7, 13729 (2016). Copyright 2016 The Author(s). (d) A design with crossed layers of V-corrugated high-density polyethylene (HDPE) foils. Reproduced with permission from Nilsson et al., Sol. Energy Mater. 12, 327–333 (1985). Copyright 1985 Elsevier. (e) A novel wind shield that can reduce the convection loss through the separation of the main airflow from the radiative cooling surface. Reproduced with permission from A. Golaka and R. H. B. Exell, Renewable Energy 32, 593–608 (2007). Copyright 2006 Elsevier. (f) Image of a UV stabilized high density polyethylene (HDPE) mesh. It has a thread diameter of 0.15 mm and a hole size of approximately 0.8 mm × 0.4 mm. The HDPE mesh structure is stronger and has a longer life than an impermeable PE film. Reproduced with permission from Gentle et al., Sol. Energy Mater. Sol. Cells 115, 79–85 (2013). Copyright 2013 Elsevier.

FIG. 13.

Convection shield to suppress parasitic thermal losses through convection and conduction to the ambient from a radiative sky cooling device. (a) Net cooling power plotted as a function of subambient temperatures under a variety of overall heat transfer coefficients. The 30-μm-thick PVDF film is employed as a radiative cooling surface. Deep subambient temperature ≥25 °C can be achieved if h is controlled to be ≤1 W/(m2 K). Results are obtained at ambient temperature 20 °C and precipitable water 10 mm. (b) Transmittance of the polyethylene (PE) film with different thicknesses in the wavelength range between 5 μm and 50 μm. Data adapted with permission from J. Appl. Phys. 52, 4205 (1981). Copyright 1981 American Institute of Physics, Nilsson et al., Sol. Energy Mater. 12, 327–333 (1985). Copyright 1985 Elsevier, Berdahl et al., Int. J. Heat Mass Transfer 26(6), 871–880 (1983). Copyright 1983 Pergamon Press Ltd. (c) Transmittance of the PE film, ZnS, and ZnSe in the infrared range. The solid red curve in (b) corresponds to the red solid curve in (c). Data adapted with permission from J. Appl. Phys. 52, 4205 (1981). Copyright 1981 American Institute of Physics, S. N. Bathgate and S. G. Bosi, Sol. Energy Mater. Sol. Cells 95, 2778–2785 (2011). Copyright 2011 Elsevier, Chen et al., Nat. Commun. 7, 13729 (2016). Copyright 2016 The Author(s). (d) A design with crossed layers of V-corrugated high-density polyethylene (HDPE) foils. Reproduced with permission from Nilsson et al., Sol. Energy Mater. 12, 327–333 (1985). Copyright 1985 Elsevier. (e) A novel wind shield that can reduce the convection loss through the separation of the main airflow from the radiative cooling surface. Reproduced with permission from A. Golaka and R. H. B. Exell, Renewable Energy 32, 593–608 (2007). Copyright 2006 Elsevier. (f) Image of a UV stabilized high density polyethylene (HDPE) mesh. It has a thread diameter of 0.15 mm and a hole size of approximately 0.8 mm × 0.4 mm. The HDPE mesh structure is stronger and has a longer life than an impermeable PE film. Reproduced with permission from Gentle et al., Sol. Energy Mater. Sol. Cells 115, 79–85 (2013). Copyright 2013 Elsevier.

Close modal

Figure 13(b) shows the transmittance of PE films at different thicknesses. Clearly, the thinner the PE, the higher the transmittance, which results in a contradictory requirement for the PE film as it has to be thin to be transparent enough in the infrared, and also has to withstand unusual environmental conditions such as wind and hail.156 Therefore, although almost ideal in terms of infrared transparency, there are concerns in using PE films due to durability and mechanical sustainability.157 For a relatively large radiative cooling aperture area (in meter scale), a thin PE film will experience up and down movement when the wind blows across the surface, causing the centimeter-thick air layer trapped between the convection shield and the radiative cooling surface to mix and transfer heat through convection. In addition, the PE film can deteriorate under solar ultraviolet (UV) radiation.158 To make the convection shield resist wind and possible hail, instead of only one layer of PE film, researcher have proposed multiple layers to achieve better performance (double layer,159 or even triple layers160).

Researchers have explored several other materials for convection shields. Figure 13(c) shows the transmittance of two other possible candidates in the wavelength range between 5 μm and 50 μm, zinc sulfide (ZnS), and zinc selenide (ZnSe).144,161 ZnSe has very high transmittance in the atmospheric window, and is mechanically stronger than PE, but much more expensive. The ZnS film is also strong and impervious to damage by solar UV,152,162 but infrared transmittance is not as good as PE and ZnSe. Other possible materials include transparent fluorinated ethylene propylene (FEP)158 and potassium chloride (KCl),93 but KCl transparent is very brittle and can be dissolved in water.93 

Usually the convection shield is a planar structure, as shown in Fig. 9(c), but some other structures have also been investigated. Nilsson et al.163 introduced a design with crossed layers of V-corrugated high-density PE films, as shown in Fig. 13(d). Typical results show that the infrared transmittance over the atmospheric window is up to 0.73 together with a thermal resistance of 1.1 m2 K W−1. Golaka and Exell156 introduced a novel wind shield that can reduce the convection loss through the separation of the main airflow from the radiative cooling surface, as depicted in Fig. 13(e). The simulation results show that the structure can effectively reduce convection loss at wind velocities above 1.5 m/s. Gentle et al.164 introduced a self-supporting PE mesh as a durable infrared transparent convection shield [Fig. 13(f)].

6. Subambient temperature reduction vs net cooling power

With the nonradiative thermal loss significantly suppressed, radiative sky cooling could even achieve cryogenic temperature in deep space,165 or in a dedicated man-made vacuum (to deep subfreezing temperatures), as shown in Fig. 9(d).144 In real-world applications, it is the requirement from end-users that determines the target temperature. Radiative sky cooling applications can involve cooling an object to a temperature as low as possible (when Pnet=0), or continuous cooling at an acceptable net cooling power (Pnet>0). Figure 14 shows the net cooling power (Pnet) calculated using Eq. (2) as a function of the targeted subambient temperature reduction at three different ambient temperatures for both daytime and nighttime. Essentially, a larger subambient temperature reduction results in a larger thermal loss to the ambient due to nonradiative heat transfer, and therefore, a smaller net cooling power. At a larger subambient temperature, the net cooling power drops at a faster rate. On the other hand, a smaller temperature difference might not generate sufficient useful cooling utility for practical applications. There is apparently a need to strategize the operation in radiative cooling systems to balance the net cooling power and the subambient temperature which are dependent on each other.

FIG. 14.

Modeling results show the net cooling power (Pnet) calculated using Eq. (2) as a function of the targeted subambient temperature reduction at three different ambient temperatures for both day and night. Nonradiative heat transfer coefficient is kept constant at 3 W/(m2 K). Precipitable water is 10 mm. Solar irradiance is kept at 700 W/m2 for daytime modeling.

FIG. 14.

Modeling results show the net cooling power (Pnet) calculated using Eq. (2) as a function of the targeted subambient temperature reduction at three different ambient temperatures for both day and night. Nonradiative heat transfer coefficient is kept constant at 3 W/(m2 K). Precipitable water is 10 mm. Solar irradiance is kept at 700 W/m2 for daytime modeling.

Close modal

7. Structure design to enhance radiative sky cooling

As introduced above, both surface emissivity and atmospheric emissivity have angular dependence. For radiative sky cooling, it is more efficient to let thermal radiation take place predominately in those angles that have large surface emissivity and small atmospheric emissivity, e.g., the zenith direction, due to the shortest atmospheric path length.30,166 In addition, though the radiative cooling surface usually has the maximum radiation intensity occurs at its normal direction [see Fig. 8(b)], a good radiative cooling surface should have emissivity variation as small as possible from 0° to 90° zenith angles.54 The angle dependence of the emissivity can be tuned at both the material level167,168 and the device level.144Figure 15 shows some structures that can increase radiative sky cooling power by blocking atmospheric radiation from large zenith angles. Jacobs et al.169 experimentally showed that by changing a radiative cooling surface from a planar structure to inverted pyramid structure, as shown in Fig. 15(b), 20% increase in cooling power is achieved. Chen et al.144 used a highly reflective inverted mirror cone placed on top of the radiative cooling surface to restrict the angular range of the apparatus to around the zenith direction where the sky is most transparent, as shown in Fig. 15(c).

FIG. 15.

Structures that can increase radiative sky cooling power by blocking atmospheric radiation from large zenith angles. The bottom areas highlighted in gray are radiative cooling surfaces. The structures that surrounding the radiative cooling surfaces are highly reflective: (a) parabolic,53 (b) inverted pyramid,169 and (c) inverted cone structures.144 

FIG. 15.

Structures that can increase radiative sky cooling power by blocking atmospheric radiation from large zenith angles. The bottom areas highlighted in gray are radiative cooling surfaces. The structures that surrounding the radiative cooling surfaces are highly reflective: (a) parabolic,53 (b) inverted pyramid,169 and (c) inverted cone structures.144 

Close modal

In this section, we summarize the daytime and nighttime radiative-cooling materials with an emphasis on the recent progresses on daytime radiative sky cooling materials.

Achieving subambient radiative cooling at nighttime is straightforward. Broadband high emissivity in the infrared range can be easily obtained (orange dashed line in Fig. 9). The majority of research works on nighttime radiative cooling materials have been focused on selective surfaces within the atmospheric window (blue solid line in Fig. 9) to achieve larger subambient temperatures. These materials can be generally categorized into the following groups as discussed below: (1) polymer-based materials [polyvinyl chloride (PVC),170 polymethyl methacrylate (PMMA),171 modified polyphenylenoxid (PPO) resin,37 composite polymer materials,142 and paints based on a polymeric binder with various pigments92,172]; (2) inorganic thin films of coating materials [silicon monoxide (SiO),54 silicon dioxide (SiO2),173 silicon oxynitride (SiOxNy),174–176 SiN,173 white pigmented paints172]; (3) gas slab [ammonia (NH3),74 ethylene (C2H4),160 ethylene oxide (C2H4O),97 and mixtures of those].

Polymer-based radiative cooling materials generally consist of a thin layer of polymer and a reflective metal, e.g., aluminum. For example, Catalanotti et al.55 developed a 12.5-μm-thick polyvinyl-fluoride (PVF) film with aluminum back coating that has strong emissivity within the atmospheric window and high infrared reflectivity out of this region, as shown in Fig. 16(a). This radiative cooler demonstrated 15 °C subambient temperature cooling when it was placed in a thermal-insulated box covered with an infrared transparent PE film. Other polymer films, including polymethylpentene (PMP, more commonly called TPX),177 and polyphenylenoxid (PPO),37 also demonstrated their capability of nighttime cooling.170,177 Another type of polymer-based material is composite materials that mix infrared transmitting polymers with nanoparticles (NPs). These nanoparticles usually have narrow absorption bands lying within the atmospheric window. By adjusting the nanoparticle concentration, the composite material can have absorption spectra lie almost entirely within the atmospheric window. Gentle and Smith142 developed a composite material that uses amorphous SiC nanoparticles as dopants (20% concentration) into 25 μm thick free-standing polyethylene, as shown in Fig. 16(b). Compared to pure polyethylene film that is highly transparent in the atmospheric window, the sharp spectral tuning by phonon resonant nanoparticles yields much larger absorption in the atmospheric window. Other nanoparticles, such as Si2N2O,178,179 SiO2,142 as well as other carbon-based nanomaterials,180 such as nanodiamonds, multiwall carbon nanotubes, and carbon black, have also been used.

FIG. 16.

Infrared transmission/reflection spectrum of some nighttime radiative cooling materials. (a) Metallized polyvinyl-fluoride thin film.55 Adapted with permission from Catalanotti et al. Solar Energy 17, 83–89 (1975). Copyright 1975 Pergamon Press. (b) Composite of a polyethylene (PE) thin film with amorphous SiC nanoparticles.142 Adapted with permission from A. R. Gentle and G. B. Smith, Nano Lett. 10, 373–379 (2010). Copyright 2010 American Chemical Society. (c) Silicon oxide and nitride film on aluminum. Adapted with permission from Eriksson et al., Sol. Energy Mater. 12, 319–325 (1985). Copyright 1985 Elsevier. (d) A slab of selectively infrared-emitting ethylene gases.160 Adapted with permission from Appl. Phys. Lett. 39(6), 507–509. Copyright 1981 American Institute of Physics.

FIG. 16.

Infrared transmission/reflection spectrum of some nighttime radiative cooling materials. (a) Metallized polyvinyl-fluoride thin film.55 Adapted with permission from Catalanotti et al. Solar Energy 17, 83–89 (1975). Copyright 1975 Pergamon Press. (b) Composite of a polyethylene (PE) thin film with amorphous SiC nanoparticles.142 Adapted with permission from A. R. Gentle and G. B. Smith, Nano Lett. 10, 373–379 (2010). Copyright 2010 American Chemical Society. (c) Silicon oxide and nitride film on aluminum. Adapted with permission from Eriksson et al., Sol. Energy Mater. 12, 319–325 (1985). Copyright 1985 Elsevier. (d) A slab of selectively infrared-emitting ethylene gases.160 Adapted with permission from Appl. Phys. Lett. 39(6), 507–509. Copyright 1981 American Institute of Physics.

Close modal

Inorganic thin films of silicon-based oxide (SiO), nitride (Si3N4), and oxynitride (SiOxNy) have been shown to have a high mid-infrared emissivity because Si-O and Si-N bonds are absorptive at 9.5 μm and 11.5 μm, respectively.54,174,181 In order to improve the emissivity, the SiO film was optimized to be 1-μm-thick to reduce infrared reflection via destructive interference.16 Bilayer and multiple layer coatings of SiO, Si3N4, and SiOxNy feature broadband infrared emissivity compared to a single SiO film.173,175Figure 16(c) shows an example of computed spectral reflectance for a bilayer SiO2 and SiO0.25N1.52 on aluminum in the wavelengths between 5 μm and 25 μm. Besides the silicon-based thin film, ceramic materials, such as MgO and LiF, also exhibit radiative sky cooling capability when they were backed with an infrared reflective layer.17 The 1.1-mm-thick MgO film has demonstrated around 20 °C subambient cooling temperature during a dry and clear sky night.17 

Spectrally selective infrared emissivity can also be achieved by using a slab of selectively infrared-emitting gas enclosed in an infrared transparent container. Some molecular stretching and rotation modes feature strong infrared absorption within the atmospheric window. For example, the C=C bond has strong light absorption between 10 μm and 12.5 μm, C-O and C-N stretching bonds are absorptive from 8 μm to 10 μm, deformation vibration of OH group provides IR absorption between 8 μm and 10 μm.28,97 Granqvist and co-authors studied the possibility of using different confined gas slabs as radiative sky cooling materials, such as NH3,74 C2H4,160 C2H4O,97 or mixtures of them. With thicknesses of a few centimeters, these gases were reported as good candidates for radiative sky cooling by characterizing their optical transmission spectrum in the mid-infrared region, as shown in Fig. 16(d) for C2H4. Field temperature measurement showed 12–18 °C subambient cooling, and subambient cooling is achieved during the day when the sunlight is shaded.28,74,160 The significant advantage to use such gas over a solid radiative cooling surface is that no extra heat transfer fluid is needed to build a cooling device.74 

Compared to nighttime cooling, daytime radiative sky cooling is much more desirable as it matches better with cooling load profile because most end-users, e.g., buildings, have their peak cooling load during the day. However, for either above-ambient (black solid spectrum line in Fig. 9) or subambient (red dashed spectrum line in Fig. 9) cooling, significant challenge exists to achieve high solar reflectance simultaneously with high emissivity over the atmospheric window. The stringent requirement seems a big challenge in the past few decades,152,182,183 with a few attempts in the 1980s and 1990s.152,184 The materials being sought after are primarily oxides and carbonates of titanium, aluminum, calcium, and zinc, due to their high reflectivity in the solar spectrum. By testing different pigments, Nilsson and Niklasson184 found that a ZnS pigmented polyethylene (PE) can achieve a solar reflectance of 0.825,152 but cannot achieve noon-time radiative sky cooling.184 Some other researchers also investigated radiative sky cooling during the day. However, the radiative cooling surface was shielded from direct solar radiation, with only diffuse solar radiation falls upon it.153,160,185 Since it is not practical to shade the radiative cooling surface from the sunlight in many scenarios, especially for large surfaces. This section focuses on the materials that show a noticeable cooling effect under direct sunlight.

One should not miss the daytime radiative sky cooling effect recently observed in many nature beings. Figure 17 shows reflectivity and absorptivity of some nanostructured biomaterials. Figures 17(a) and 17(b) show that the Saharan silver ant's body are covered by a dense array of hair that is strongly reflective in solar wavelengths between 0.4 μm and 1.7 μm, and emissive in mid-infrared wavelengths between 2.5 μm and 16 μm to maintain a comfortable body temperature in a hot desert.186Figure 17(c) shows a Bistonina biston butterfly's scent pad/patch with a unique nanostructure that is very emissive in the mid-infrared (2.5–16 μm) to dissipate heat and has high reflectivity to reduce solar absorption, enabling it to cool the wings.187 Another example in nature is the cocoon made by wild silk moths, which is a random structure of fibers, exhibiting high solar reflectivity due to the scattering of light along with high mid-IR emissivity. It can protect moth pupae from overheating under the sun [see Fig. 17(d)].188 These nanostructured biomaterials could have an inspiration for the development of daytime radiative sky cooling materials.

FIG. 17.

Daytime radiative sky cooling in nanostructured biomaterials. (a) and (b) The reflectivity spectra of the Saharan silver ant's body surface with and without hair in the visible and near infrared (0.4–1.7 μm), and mid-infrared region (2.5–16 μm), respectively.186 Reflectivity = 1 − Emissivity. Reproduced with permission from Shi et al., Science 349(6245), 298–301. Copyright 2015 The American Association for the Advancement of Science. (c) Infrared absorptivity (emissivity) spectra of scent pad (red), black region (black) and blue region (blue) of Bistonina biston butterfly's wings.187 Reproduced with permission from Tsai et al., in 2017 Conference on Lasers and Electro-Optics (CLEO) (2017), pp. 1–2. Copyright 2017 Optical Society of America. (d) The hemispherical reflection, transmission and absorption spectra of cocoon shell from 400 nm to 14 μm, showing high solar reflectance and high thermal emittance.188 Reproduced with permission from Shi et al., in 2017 Conference on Lasers and Electro-Optics (CLEO) (2017), pp. 1–2. Copyright 2017 Optical Society of America.

FIG. 17.

Daytime radiative sky cooling in nanostructured biomaterials. (a) and (b) The reflectivity spectra of the Saharan silver ant's body surface with and without hair in the visible and near infrared (0.4–1.7 μm), and mid-infrared region (2.5–16 μm), respectively.186 Reflectivity = 1 − Emissivity. Reproduced with permission from Shi et al., Science 349(6245), 298–301. Copyright 2015 The American Association for the Advancement of Science. (c) Infrared absorptivity (emissivity) spectra of scent pad (red), black region (black) and blue region (blue) of Bistonina biston butterfly's wings.187 Reproduced with permission from Tsai et al., in 2017 Conference on Lasers and Electro-Optics (CLEO) (2017), pp. 1–2. Copyright 2017 Optical Society of America. (d) The hemispherical reflection, transmission and absorption spectra of cocoon shell from 400 nm to 14 μm, showing high solar reflectance and high thermal emittance.188 Reproduced with permission from Shi et al., in 2017 Conference on Lasers and Electro-Optics (CLEO) (2017), pp. 1–2. Copyright 2017 Optical Society of America.

Close modal

Nanophotonic structures with structural features at a wavelength or subwavelength scale, can have distinct thermal radiation properties.189 In 2013, Rephaeli et al.78 introduced a metal-dielectric nanophotonic structure and predicted theoretically the possibility to achieve daytime radiative sky cooling. The photonic structure is a two-dimensional (2D) periodic pillar array consisting of quartz and SiC as top two layers on a stack of high- and low-index dielectric materials on a silver substrate, as shown in Fig. 18(a). The emissivity is shown to be enhanced across the whole atmospheric window by phonon-polariton resonance from quartz and the SiC layer, with a net cooling power greater than 100 W/m2 achievable at ambient temperature under the sun. In 2014, Raman et al.31 experimentally demonstrated subambient daytime radiative sky cooling with an integrated photonic thermal radiative cooler. This radiative cooler consists of seven layers of silicon dioxide (SiO2) and hafnium dioxide (HfO2) with optimized thickness on top of the 200 nm silver (Ag) coated silicon wafer [Fig. 18(b)]. The bottom four layers of SiO2 and HfO2 are thinner and primarily responsible for solar reflection, while the top three layers are much thicker and designed for thermal radiation. The structure can reflect 97% of incident sunlight and emitting strongly in the atmospheric window [Figs. 18(c) and 18(d)]. With an apparatus that minimizes nonradiative heat losses on the radiative cooler, a 4.9 °C subambient temperature difference and 40.1 ± 4.1 W/m2 cooling power were achieved when the photonic radiator is exposed to direct sunlight exceeding 850 W/m2 [Figs. 18(e) and 18(f)].

FIG. 18.

First experimental demonstration of daytime subambient radiative sky cooling by using a nanophotonic radiative cooler. (a) Absorptivity/emissivity spectra of 2D the periodic pillar array show high solar reflection and infrared emission in the atmospheric window.78 Reproduced with permission from Rephaeli et al., Nano Lett. 13, 1457–1461 (2013). Copyright 2013 American Chemical Society. (b) The photonic radiative cooler is a multilayer coating structure of SiO2 and HfO2 on the Ag coated substrate. (c) Spectral emissivity of the photonic radiative cooler in solar spectra region. (d) Spectral emissivity of the photonic radiative cooler in mid-infrared region. (e) Subambient cooling demonstrated during the day with a maximum temperature difference of 4.9 °C. (f) The measured cooling powers at a various of subambient temperatures.31 Figures (b)–(f) are reproduced with permission from Raman et al., Nature 515, 540 (2014). Copyright 2014 Macmillan Publishers Limited.

FIG. 18.

First experimental demonstration of daytime subambient radiative sky cooling by using a nanophotonic radiative cooler. (a) Absorptivity/emissivity spectra of 2D the periodic pillar array show high solar reflection and infrared emission in the atmospheric window.78 Reproduced with permission from Rephaeli et al., Nano Lett. 13, 1457–1461 (2013). Copyright 2013 American Chemical Society. (b) The photonic radiative cooler is a multilayer coating structure of SiO2 and HfO2 on the Ag coated substrate. (c) Spectral emissivity of the photonic radiative cooler in solar spectra region. (d) Spectral emissivity of the photonic radiative cooler in mid-infrared region. (e) Subambient cooling demonstrated during the day with a maximum temperature difference of 4.9 °C. (f) The measured cooling powers at a various of subambient temperatures.31 Figures (b)–(f) are reproduced with permission from Raman et al., Nature 515, 540 (2014). Copyright 2014 Macmillan Publishers Limited.

Close modal

A few other nanophotonic structures, such as metal-dielectric multilayer conical pillar arrays,190 metal-loaded dielectric resonator metasurfaces,191 metal-dielectric-metal resonators,192 multilayer pyramidal nanostructures,193 and nanoparticle embedded double-layer coatings194 have also shown high reflectivity in the solar spectrum and high emissivity over the atmospheric window. Hossain et al.190 introduced a design of a thermal emitter with multilayer conical metamaterial pillar arrays, as shown in Fig. 19(a). Each pillar consists of alternating layers of aluminum and germanium. The thickness of the aluminum layer is 30 nm and the thickness of the germanium layer is 110 nm. Figure 19(b) shows that the structure can achieve very high emissivity within the entire atmospheric window (8–13 μm) and low emissivity outside the atmospheric window. With a solar reflector that can minimize the total solar power absorption down to 3%, the emitter can possess a radiative cooling power of 116.6 W/m2 at ambient temperature. Zou et al.191 introduced a metal-loaded dielectric resonator metasurface for radiative sky cooling. The designed surface consists of phosphorous-doped n-type silicon and silver, as shown by a false-color scanning electron microscope image in Fig. 19(c). The spectral absorptivity measurement showed a strong absorption/emission band that matches well with the atmospheric window, as shown in Fig. 19(d). This metasurface has the advantage of convenient integration with various silicon-based photonic/electronic platforms. Wu et al.193 introduced a multilayer all-dielectric micropyramid structure to achieve high mid-infrared selectivity in planar photonic devices, as shown in Fig. 19(e). The optimized structure achieves broadband selective emissivity in the atmospheric window and extremely low absorption in the entire solar spectrum, as shown in Fig. 19(f).

FIG. 19.

Nanophotonic structures for daytime radiative sky cooling. (a) Radiative emitter with multilayer conical metamaterial pillar arrays. (b) Absorptivity/emissivity of the multilayer conical metamaterial pillar arrays between 5 μm and 17 μm.190 Adapted with permission from Hossain et al., Adv. Opt. Mater. 3, 1047–1051 (2015). Copyright 2015 Wiley. (c) Radiative cooler with a metal-loaded dielectric resonator metasurface. (d) Absorptivity/emissivity of the metal-loaded dielectric resonator metasurface.191 Adapted with permission from Zou et al. Adv. Opt. Mater. 5, 1700460 (2017). Copyright 2017 Wiley. (e) Radiative cooling with a multilayer all-dielectric micropyramid structure. (f) Absorptivity/emissivity of the multilayer all-dielectric micropyramid structure.193 Adapted with permission from Wu et al., Mater. Des. 139, 104–111 (2018). Copyright 2017 Elsevier.

FIG. 19.

Nanophotonic structures for daytime radiative sky cooling. (a) Radiative emitter with multilayer conical metamaterial pillar arrays. (b) Absorptivity/emissivity of the multilayer conical metamaterial pillar arrays between 5 μm and 17 μm.190 Adapted with permission from Hossain et al., Adv. Opt. Mater. 3, 1047–1051 (2015). Copyright 2015 Wiley. (c) Radiative cooler with a metal-loaded dielectric resonator metasurface. (d) Absorptivity/emissivity of the metal-loaded dielectric resonator metasurface.191 Adapted with permission from Zou et al. Adv. Opt. Mater. 5, 1700460 (2017). Copyright 2017 Wiley. (e) Radiative cooling with a multilayer all-dielectric micropyramid structure. (f) Absorptivity/emissivity of the multilayer all-dielectric micropyramid structure.193 Adapted with permission from Wu et al., Mater. Des. 139, 104–111 (2018). Copyright 2017 Elsevier.

Close modal

Cost and scalability is always of interest for real-world energy solutions. In 2017, Zhai et al.32,195,196 reported a ground-breaking approach to make a metamaterial that is scalable and low cost, as shown in Fig. 20. The “designer metamaterial” consists of a transparent TPX polymer over the solar spectra, encapsulating randomly distributed silicon dioxide (SiO2) microspheres and backed by a 200-nm-thick silver layer [Fig. 20(a)]. The metamaterial is extremely emissive across the entire atmospheric window (8–13 μm) due to phonon-enhanced Fröhlich resonances of the embedded microspheres [Fig. 20(b)]. Such a glass-polymer hybrid metamaterial thin film can generate 93 W/m2 radiative cooling power at noon under direct sunlight, and an average of 110 W/m2 radiative cooling power for three consecutive days, as shown in Fig. 20(c). The metamaterial can be scalably-manufactured by the mature roll-to-roll extrusion and web coating systems, which makes large-scale application of radiative sky cooling possible [Figs. 20(d) and 20(e)].

FIG. 20.

Scalable-manufactured daytime radiative sky cooling metamaterial developed at the University of Colorado Boulder.32,196 (a) Schematic of the glass-polymer hybrid metamaterial backed with a thin silver film. The silver film reflects most of the incident solar irradiance and is highly emissive in the atmospheric window. (b) The full spectroscopic performance of the glass-polymer hybrid metamaterial. The metamaterial reflects ∼96% solar irradiation while possessing emissivity >0.93 in the atmospheric window. (c) The continuous measurement of cooling power over three days shows an average cooling power >110 W/m2 and a noon-time cooling power of 93 W/m2 between 11 am–2 pm. (d) A photo showing the 300-mm-wide hybrid metamaterial thin film that was produced in a roll-to-roll manner, at a speed of 5 m/min. The film is 50 μm in thickness and not yet coated with silver. (e) A photo showing the 300-mm-wide hybrid metamaterial thin film that has been coated with silver by roll-to-roll web coating. Figures (a)–(d) are reproduced with permission from Zhai et al., Science 355, 1062–1066 (2017). Copyright 2017 The American Association for the Advancement of Science.

FIG. 20.

Scalable-manufactured daytime radiative sky cooling metamaterial developed at the University of Colorado Boulder.32,196 (a) Schematic of the glass-polymer hybrid metamaterial backed with a thin silver film. The silver film reflects most of the incident solar irradiance and is highly emissive in the atmospheric window. (b) The full spectroscopic performance of the glass-polymer hybrid metamaterial. The metamaterial reflects ∼96% solar irradiation while possessing emissivity >0.93 in the atmospheric window. (c) The continuous measurement of cooling power over three days shows an average cooling power >110 W/m2 and a noon-time cooling power of 93 W/m2 between 11 am–2 pm. (d) A photo showing the 300-mm-wide hybrid metamaterial thin film that was produced in a roll-to-roll manner, at a speed of 5 m/min. The film is 50 μm in thickness and not yet coated with silver. (e) A photo showing the 300-mm-wide hybrid metamaterial thin film that has been coated with silver by roll-to-roll web coating. Figures (a)–(d) are reproduced with permission from Zhai et al., Science 355, 1062–1066 (2017). Copyright 2017 The American Association for the Advancement of Science.

Close modal

The pioneering studies by the Stanford group31 and by the Colorado group32 have inspired many researchers and have revived the interest in radiative sky cooling. Table II lists some successful experimental demonstration of daytime subambient radiative sky cooling under direct sunlight. Apparently manufacturing scalability is now an important factor, among which polymers and paints show great advantages.58,194,197–202 By coating a commercially available polyester reflector with a 200-nm-thick silver layer [Fig. 21(b)], Gentle and Smith143 successfully demonstrated a 2 °C subambient cooling under the sun without using a convection shield. The so-called “supercool roof” can be manufactured on a large scale. However, the authors mentioned that refinements are still needed to make the reflectance diffuse to avoid glare. By using a dual-layer structure consisting of a polytetrafluoroethylene (PTFE) sheet on top of a silver film [Fig. 21(c)], Yang et al.203 reported that a record-high solar reflectance (99%) has been achieved from spectrum analysis, as shown in Fig. 21(a).

TABLE II.

List of successful demonstration of daytime radiative sky cooling under direct sunlight in field thermal measurement.

MaterialsFabrication ProcessCooling performanceYearLocationReference
Seven layers of alternating silicon dioxide (SiO2) and hafnium dioxide (HfO2) thermal emitter backed by a silver (Ag) solar reflector The thermal emitter was deposited using electron beam evaporation on top of a silicon wafer. A 20-nm-thick adhesion layer of Ti was first deposited, followed by 200 nm of Ag, and the seven alternating dielectric layers of HfO2 and SiO2 at desired thicknesses Producing cooling power of 40 W/m2 and a temperature reduction of nearly 5 °C below ambient 2014 Stanford, California Raman et al.31  
Vikuiti Enhanced Specular Reflector (ESR) film with a 200-nm-thick silver coating at the bottom The material was fabricated by adding a silver coating to the Vikuiti Enhanced Specular Reflector (ESR) film. The coating method was not mentioned Achieving 2 °C subambient cooling under 1060 W/m2 solar irradiance with no convection shield 2015 Sydney, Australia Gentle and Smith143 
Randomized SiO2 microparticles dispersed in TPX polymers The selectively emissive polymer layer was fabricated through roll-to-roll extrusion, followed by roll-to-roll deposition of the reflective metal layer via plasma-enhanced magnetron sputtering Average cooling power of 93 W/m2 at noon and >110 W/m2 over a continuous 72-h day-/nighttime measurement 2017 Cave Creek, Arizona Zhai et al.32  
A fused silica coated by polydimethylsiloxane (PDMS) and Ag on the top and bottom sides, respectively The silver film was deposited by electron beam evaporation under high vacuum. The PDMS film was spin-coated for 60 s followed by degassing for 10 min and curing for 1 h at 80 °C Temperature reduction of 8.2 °C at noon time under direct sunlight 2017 Pasadena, California Kou et al.202  
A visibly-reflective extruded copolymer mirror (3 M Vikiuiti ESR film), on top of an enhanced silver reflective surface Visibly-reflective extruded copolymer mirror (3M Vikiuiti ESR film) was placed on top of an enhanced silver reflective surface Cool water up to 5 °C below the ambient air temperature at water flow rates of 0.2 l min−1 m−2, corresponding to an effective heat rejection flux of up to 70 W/m2 2017 Stanford, California Goldstein et al.51  
Hierarchically porous poly(vinylidene fluoride-co-hexafluoropropene) (P(VdF-HFP)HPPhase-inversion-based method was used to prepare a precursor solution of P(VdF-HFP) (polymer) and water (nonsolvent) in acetone (solvent) to achieve hierarchically porous polymers Subambient temperature reduction of ∼6 °C and cooling powers of ∼96 W/m2 under solar intensities of 890 and 750 W m−2, respectively 2018 Phoenix, Arizona Mandal et al.58  
MaterialsFabrication ProcessCooling performanceYearLocationReference
Seven layers of alternating silicon dioxide (SiO2) and hafnium dioxide (HfO2) thermal emitter backed by a silver (Ag) solar reflector The thermal emitter was deposited using electron beam evaporation on top of a silicon wafer. A 20-nm-thick adhesion layer of Ti was first deposited, followed by 200 nm of Ag, and the seven alternating dielectric layers of HfO2 and SiO2 at desired thicknesses Producing cooling power of 40 W/m2 and a temperature reduction of nearly 5 °C below ambient 2014 Stanford, California Raman et al.31  
Vikuiti Enhanced Specular Reflector (ESR) film with a 200-nm-thick silver coating at the bottom The material was fabricated by adding a silver coating to the Vikuiti Enhanced Specular Reflector (ESR) film. The coating method was not mentioned Achieving 2 °C subambient cooling under 1060 W/m2 solar irradiance with no convection shield 2015 Sydney, Australia Gentle and Smith143 
Randomized SiO2 microparticles dispersed in TPX polymers The selectively emissive polymer layer was fabricated through roll-to-roll extrusion, followed by roll-to-roll deposition of the reflective metal layer via plasma-enhanced magnetron sputtering Average cooling power of 93 W/m2 at noon and >110 W/m2 over a continuous 72-h day-/nighttime measurement 2017 Cave Creek, Arizona Zhai et al.32  
A fused silica coated by polydimethylsiloxane (PDMS) and Ag on the top and bottom sides, respectively The silver film was deposited by electron beam evaporation under high vacuum. The PDMS film was spin-coated for 60 s followed by degassing for 10 min and curing for 1 h at 80 °C Temperature reduction of 8.2 °C at noon time under direct sunlight 2017 Pasadena, California Kou et al.202  
A visibly-reflective extruded copolymer mirror (3 M Vikiuiti ESR film), on top of an enhanced silver reflective surface Visibly-reflective extruded copolymer mirror (3M Vikiuiti ESR film) was placed on top of an enhanced silver reflective surface Cool water up to 5 °C below the ambient air temperature at water flow rates of 0.2 l min−1 m−2, corresponding to an effective heat rejection flux of up to 70 W/m2 2017 Stanford, California Goldstein et al.51  
Hierarchically porous poly(vinylidene fluoride-co-hexafluoropropene) (P(VdF-HFP)HPPhase-inversion-based method was used to prepare a precursor solution of P(VdF-HFP) (polymer) and water (nonsolvent) in acetone (solvent) to achieve hierarchically porous polymers Subambient temperature reduction of ∼6 °C and cooling powers of ∼96 W/m2 under solar intensities of 890 and 750 W m−2, respectively 2018 Phoenix, Arizona Mandal et al.58  
FIG. 21.

Highlight of some polymeric daytime radiative sky cooling materials that can potentially be scalable-manufactured. (a) Measured emissivity/absorptivity of some daytime radiative sky cooling materials recently developed. The normalized solar irradiation of AM1.5 spectrum and clear sky transmittance in the infrared range are plotted for visualization. Adapted with permission from Bao et al., Sol. Energy Mater. Sol. Cells 168, 78–84 (2017). Copyright 2017 Elsevier. Atiganyanun et al., ACS Photonics 5, 1181–1187 (2018). Copyright 2018 American Chemical Society. Yang et al., Sol. Energy 169, 316–324 (2018). Copyright 2018 Elsevier. A. R. Gentle and G. B. Smith, Adv. Sci. 2, 1500119 (2015). Copyright 2015 WILEY. (b) A polymer mirror that consists of 300 layers of Polyethylene terephthalate (PET)/ECDEL pairs. ECDEL is a copolyester using 1,4-cyclohexane dicarboxylic acid, 1,4-cyclohexane dimethanol, and polytetramethylene ether glycol. The bottom Ag layer is added to raise solar reflectance in the near-infrared region. (c) A dual-layer structure that consists of a PTFE sheet and a back reflector made of the Ag film. Note that the Ag film is coated on a substrate and there exists an air gap between the PTFE sheet and the Ag film. (d) Double layer coating composed of TiO2 and SiO2 on an Al substrate. The densely packed TiO2 top layer is for solar reflection and the underlying SiO2 layer has selective emissivity within the atmosphere window. (e) A paint-format microsphere-based photonic random media that can potentially be scalable-manufactured.

FIG. 21.

Highlight of some polymeric daytime radiative sky cooling materials that can potentially be scalable-manufactured. (a) Measured emissivity/absorptivity of some daytime radiative sky cooling materials recently developed. The normalized solar irradiation of AM1.5 spectrum and clear sky transmittance in the infrared range are plotted for visualization. Adapted with permission from Bao et al., Sol. Energy Mater. Sol. Cells 168, 78–84 (2017). Copyright 2017 Elsevier. Atiganyanun et al., ACS Photonics 5, 1181–1187 (2018). Copyright 2018 American Chemical Society. Yang et al., Sol. Energy 169, 316–324 (2018). Copyright 2018 Elsevier. A. R. Gentle and G. B. Smith, Adv. Sci. 2, 1500119 (2015). Copyright 2015 WILEY. (b) A polymer mirror that consists of 300 layers of Polyethylene terephthalate (PET)/ECDEL pairs. ECDEL is a copolyester using 1,4-cyclohexane dicarboxylic acid, 1,4-cyclohexane dimethanol, and polytetramethylene ether glycol. The bottom Ag layer is added to raise solar reflectance in the near-infrared region. (c) A dual-layer structure that consists of a PTFE sheet and a back reflector made of the Ag film. Note that the Ag film is coated on a substrate and there exists an air gap between the PTFE sheet and the Ag film. (d) Double layer coating composed of TiO2 and SiO2 on an Al substrate. The densely packed TiO2 top layer is for solar reflection and the underlying SiO2 layer has selective emissivity within the atmosphere window. (e) A paint-format microsphere-based photonic random media that can potentially be scalable-manufactured.

Close modal

Paints and coatings can potentially be more cost-effective and applied to almost any surfaces in a relatively easy way. Bao et al.204 introduced a double-layer nanoparticle-based coatings [Fig. 21(d)], which is composed of a top solar reflection layer and an underlying layer that has high emissivity in the atmospheric window. The material has high selective emissivity in the atmospheric window, but the solar absorption at wavelengths 0.3–0.4 μm is still high, as shown in Fig. 21(a). Atiganyanun et al.199 introduced a paint-format microsphere-based photonic random media. The random media consists of silica microspheres as shown in Fig. 21(e). The size of microspheres and the filling fraction are two important parameters that determine performance. While those two parameters are not optimized, the authors showed that the material outperformed a commercial solar reflective white paint and achieved subambient cooling for most of time during the day. Recently, Mandal et al.58 introduced hierarchically porous poly(vinylidene fluoride-co-hexafluoropropene) (P(VdF-HFP)HP) coatings for passive daytime radiative sky cooling. The idea is to replace the pigments in the commercial cool-roof paints with light-scattering air voids to eliminate pigments' ultraviolet absorption. A hemispherical solar reflectance of 0.96 ± 0.03 and infrared (7–18 μm) emittances of 0.97 ± 0.02 were reported for this material, along with a subambient temperature reduction of ∼6 °C.

It is important to note that some daytime radiative cooling materials need to be scalably manufactured on an existing substrate, e.g., glass in solar cell cooling application. Lu et al.205 introduced ultrabroadband versatile glass textures that exhibit emissivity >0.96 across the atmospheric window and solar transmittance and haze above 0.94 and 0.95 at the wavelengths from 350 to 750 nm, respectively. The authors modified a silica sol-gel imprinting process to transfer the pyramidal textures from textured crystalline silicon wafers with a desirable feature size (between 1.2 and 2.5 μm) to low iron glass substrates. Figure 22 shows the modified sol-gel imprinting process. A crystalline silicon wafer with a desired feature size was used as an original master template. The original master was used to shape an intermediate PDMS mold, which was later imprinted onto a sol-gel silica layer to replicate the pyramidal textures. By applying the ultrabroadband imprinted glass to silicon PV modules as an encapsulant cover, the short-circuit current and energy conversion efficiency were increased relatively by 5.12% and 3.13%, respectively. The authors claimed that the fabrication of such textures is photolithography-free, scalable, and compatible with the PV industry.

FIG. 22.

Schematic illustration on the preparation of ultrabroadband versatile textures for radiative sky cooling and light management.205 Commercially available crystalline silicon wafer with the desired features was used as an original master. The original master was used to shape an intermediate PDMS mold, which was later imprinted onto a sol-gel silica layer to replicate the pyramidal textures. Figure reproduced with permission from Lu et al., Sol. RRL 1, 1700084 (2017). Copyright 2017 Wiley.

FIG. 22.

Schematic illustration on the preparation of ultrabroadband versatile textures for radiative sky cooling and light management.205 Commercially available crystalline silicon wafer with the desired features was used as an original master. The original master was used to shape an intermediate PDMS mold, which was later imprinted onto a sol-gel silica layer to replicate the pyramidal textures. Figure reproduced with permission from Lu et al., Sol. RRL 1, 1700084 (2017). Copyright 2017 Wiley.

Close modal

Polymer-based radiative sky cooling metamaterials are likely to be commonly used in the future due to their significant advantage over other nonpolymer nanophotonic materials in terms of manufacturability and cost.206 However, several challenges still remain unsolved with reliability being the most critical one. Maintaining high infrared emissivity in the atmospheric window may not be a challenge. However, maintaining high solar reflectivity turns out to be much more difficult due to the degradation of the polymer itself and the metal solar-reflective layer, e.g., the silver layer in Figs. 21(b) and 21(c). Oxidation of the metal layer by oxygen and moisture permeating from the front and back surfaces causes a major threat to the long-term solar reflectivity. Several approaches can be taken to tackle this problem. For example, increasing the thickness of the emissive polymer layer and/or the metal layer slows and delays oxygen and moisture permeation, while using a barrier polymer, such as polyvinylidene chloride (PVDC) and polychlorotrifluoroethylene (PCTFE), against water and oxygen, can be another solution. Alternatively, a thin layer of barrier material can be coated onto the polymer and metal layers. It is worth noting that the recent study by Mandal et al.58 shows that it is possible to achieve high solar reflectivity without backing the polymer layer with a reflective metal layer. Through a phase inversion method, micro- and nanoscale pores are generated in polymers, which strongly scatter and reflect light in all directions. Although this type of radiative sky cooling polymers without the metal layer could be a good alternative to metal-backed radiative sky cooling polymeric metamaterials, the reliability needs to be thoroughly investigated, for example, whether the porous structures are prone to dust and moisture accumulation. Another aspect of long-term reliability is the prevention of the polymer layer from UV and chemical degradation. UV degradation of polymers is a well-known phenomenon. There are polymers such as PET and PTFE that are not easily affected by UV. Anti-UV additives may be added to increase the lifetime.

Integrating functionality with radiative sky cooling materials is also of interest. A great example is the demonstration of self-adaptive/switchable radiative sky cooling based on phase change materials for active temperature control.207–209 Since ambient temperature varies during the day and throughout the year, radiative sky cooling may not needed at night and in the winter season. It is thus desirable to design a switchable radiative cooling material that can adapt itself according to different ambient temperatures. Ono et al.207 designed a photonic structure to realize self-adaptive radiative cooling. The structure consists of a spectrally-selective filter (11 layers of Ge/MgF2) as the top layer and a three-layer VO2/MgF2/W structure as the bottom radiative cooler, as shown in Fig. 23(a). When VO2 experiences phase transition from the insulating to metallic state, absorptivity/emissivity does not have too much change in the solar spectrum as shown in Fig. 23(b), but there is significant change in the mid-infrared wavelength range (8–13 μm) as shown in Fig. 23(c). Another study by Wu et al.208 proposed a structure consisting of a VO2/SO2 multilayer metasurface for radiative sky cooling as shown in Fig. 23(d), where a distinct change in cooling power can be seen when experiencing a phase-change [Fig. 23(e)].

FIG. 23.

Photonic structures to realize self-adaptive/switchable radiative sky cooling. (a) A two-layer photonic structure consists of a spectrally-selective filter (11 layers of Ge/MgF2) as the top layer and a three layer VO2/MgF2/W structure as the bottom radiative cooler.207 (b) Solar (0.3–2.5 μm) absorptivity/emissivity of the photonic structure in metallic and insulating states. (c) Mid-infrared (2.5–15 μm) absorptivity/emissivity of the photonic structure in metallic and insulating states. Reproduced with permission from Ono et al., Opt. Express 26(18), A777–A787 (2018). Copyright 2018 The Optical Society. (d) A structure consisting of a VO2/SO2 multilayer metasurface for passive temperature control. (e) Dramatic cooling power change after phase transition. Tdev is the temperature difference between the structure and ambient temperature, and ambient temperature is assumed to be 27 °C. Reproduced with permission from Wu et al., Sci. Rep. 8, 7684 (2018). Copyright 2018 Nature Publishing Group.

FIG. 23.

Photonic structures to realize self-adaptive/switchable radiative sky cooling. (a) A two-layer photonic structure consists of a spectrally-selective filter (11 layers of Ge/MgF2) as the top layer and a three layer VO2/MgF2/W structure as the bottom radiative cooler.207 (b) Solar (0.3–2.5 μm) absorptivity/emissivity of the photonic structure in metallic and insulating states. (c) Mid-infrared (2.5–15 μm) absorptivity/emissivity of the photonic structure in metallic and insulating states. Reproduced with permission from Ono et al., Opt. Express 26(18), A777–A787 (2018). Copyright 2018 The Optical Society. (d) A structure consisting of a VO2/SO2 multilayer metasurface for passive temperature control. (e) Dramatic cooling power change after phase transition. Tdev is the temperature difference between the structure and ambient temperature, and ambient temperature is assumed to be 27 °C. Reproduced with permission from Wu et al., Sci. Rep. 8, 7684 (2018). Copyright 2018 Nature Publishing Group.

Close modal

Though many radiative cooling materials have been developed with very appealing spectroscopic properties, not many reliable cooling systems are available yet due to another set of challenges beyond materials. The intrinsically low energy density of radiative sky cooling represents one of the most important challenges for the practical use of this technology for either large scale systems such as power plant cooling or the interest in electronics cooling which in general has a power density on the order of 100 W/cm2. Other issues that need to be considered when building radiative sky cooling devices and systems include the system configuration and controls, cooling load profiles from end-users, the impact of weather conditions, as well as the system cost and payback period.

The use of nighttime radiative sky cooling can be traced back to 400 BC when Persians made ice during a clear cold winter night, even though mechanisms of this phenomena was completely not understood yet.14,210 In 1700s, the water desalination system was built by freezing to purify sea water and impure water.211 In this section, we review existing applications of radiative sky cooling as well as those that are currently been explored. A few potential application areas, such as cooling of buildings and solar cells, dew water harvesting, and supplemental cooling for power plants are discussed in detail. The use of daytime radiative sky cooling is highlighted as it represents an exciting new direction over the last few years.

Buildings consume approximately 40% of total energy use in the United States, and most of energy consumption is for heating, ventilation, and air conditioning (HVAC) systems.212,213 More specifically, space cooling accounts for about 12% of building energy consumption in the United States.214 The integration of radiative sky cooling technology with buildings, which consumes no or very little energy, could have tremendous impact on world energy saving and CO2 emission reduction. Earlier works on radiative sky cooling in buildings were summarized in review articles by Lu et al.,34 Vall and Castell,118 and Zhao et al.120 In general, the use of radiative sky cooling can be achieved either in a passive way such as cool roof (with no energy input), or in a more active way such as cooling of a heat transfer fluid and then use the heat transfer fluid for building cooling (requiring a small percentage of energy input for pumping). Passive systems have the advantage of system simplicity, but suffer from possibly excessive cooling when cooling is not needed, especially during the winter or the nighttime. Active systems can be better controlled, and thus are more feasible for those applications where cooling power needs to be regulated, but the systems are much more complex. The efficiency of a radiative sky cooling system in a building depends on the availability of exposed flat (or low slope) roof area and the cooling potential that can be explored in single floor buildings or in the top floor of multifloor buildings.215 This section presents an overview of radiative sky cooling for buildings, and more importantly, the opportunities and challenges in deploying this technology. Note that when using the term “passive,” some researchers highlight the use of a natural cooling source and does not exclude the use of a fan or pump to enhance system performance.34,216 In this work, the term passive refers to the technology with zero energy input. Those systems with a fan or a pump are described as an active system in Sec. IV A 2.

1. Passive applications

Passive cooling of buildings with a large rooftop is an effective way to reduce building energy consumption, especially considering that the cooling load of most buildings are on the order of 100 W/m2 of building area.216,217 Nahar et al.218 experimentally compared eight different passive techniques, such as thermal insulation on roof, roof pond, and evaporative cooling for building cooling. They found that white-colored roof (i.e., “cool roof”) is the best for thermal comfort in arid areas, especially when considering the ease of operation and without consumption of water. Due to its special advantage in energy efficiency, cool roof is one of the major and prevailing passive cooling techniques identified for building applications.219 

Cool roof refers to roof materials that have been designed to have high solar reflectance and high thermal emissivity (in mid-infrared regions) than a regular roof, which can lower roof temperatures and reduce heat transfer into a building, as shown in Figs. 24(a) and 24(b). The use of cool roof to increase albedo (i.e., solar reflectance) could reduce the building cooling load, lower peak electricity demand, improve indoor thermal comfort for unconditioned space, extend roof service life, and reduce building's contribution to the urban heat island [Fig. 24(c)].222–225 The solar reflectance (or albedo) and thermal emissivity are the two important performance indexes associated with a cool roof material. Cool roof products are versatile including tiles, paintings or coatings and can be found at the website of the cool roof rating council.226 The solar reflectance and thermal emissivity values of these products are usually also available via the cool roof rating council standard (CRRC-1).226 An optimized cool roof should have optimized thermal radiative properties (albedo and thermal emissivity) and roof R-value (i.e., thermal insulation).227 One common method to quantify cool roofs is the estimation of their solar reflectance index (SRI), which is calculated by using solar reflectance and thermal emissivity values adjusted with wind coefficients. According to the active standard ASTM E1980, SRI is defined by the equation

(14)

where

(15)

Here, α is the solar absorptance, ε is the thermal emissivity, and hc is the convective heat transfer coefficient. By definition, cool roofs should have an initial SRI value of at least 78.228 Table III gives a list of regular roofing materials and cool roof materials.

FIG. 24.

Passive radiative sky cooling for buildings—cool roof application. (a) Cool roof reduces heat flow into buildings through enhanced solar reflection and infrared thermal emission. (b) A high albedo cool roof (left) compared to a galvanized low albedo roof (right). Much lower surface temperature is observed for the cool roof.220 Reproduced with permission from Smith et al., Renewable Energy Environ. Sustainable 2, 13 (2017). Copyright 2017 EDF Sciences. (c) The use of cool roof can reduce building's contribution to the urban heat island.221 Reproduced with permission from Akbari et al., Cooling our Communities. A Guidebook on Tree Planting and Light-Colored Surfacing. Copyright 1992 US Environmental Protection Agency. US Environmental Protection Agency.

FIG. 24.

Passive radiative sky cooling for buildings—cool roof application. (a) Cool roof reduces heat flow into buildings through enhanced solar reflection and infrared thermal emission. (b) A high albedo cool roof (left) compared to a galvanized low albedo roof (right). Much lower surface temperature is observed for the cool roof.220 Reproduced with permission from Smith et al., Renewable Energy Environ. Sustainable 2, 13 (2017). Copyright 2017 EDF Sciences. (c) The use of cool roof can reduce building's contribution to the urban heat island.221 Reproduced with permission from Akbari et al., Cooling our Communities. A Guidebook on Tree Planting and Light-Colored Surfacing. Copyright 1992 US Environmental Protection Agency. US Environmental Protection Agency.

Close modal
TABLE III.

List of some regular roofing materials and some cool roof material properties.

MaterialsSolar reflectanceThermal emittanceSRIReference
Regular roofing materials Generic black shingle 0.05 0.91 57  
Shasta white shingle 0.26 0.91 27 57  
Light gravel 0.34 0.90 37 57  
White coating on shingle 0.71 0.91 87 57  
Cool roof materials United Coatings™ 0.87 0.87 110 229 and 230  
Kymax™ Coating 
Miracool silicon acrylic emulsion paint 0.89 0.89 113 229 and 231  
Renolit Alkorbright polyester-reinforced roofing membrane 0.91 0.84 115 229 and 232  
APOC 256X coating 0.94 0.88 120 229 and 233  
MaterialsSolar reflectanceThermal emittanceSRIReference
Regular roofing materials Generic black shingle 0.05 0.91 57  
Shasta white shingle 0.26 0.91 27 57  
Light gravel 0.34 0.90 37 57  
White coating on shingle 0.71 0.91 87 57  
Cool roof materials United Coatings™ 0.87 0.87 110 229 and 230  
Kymax™ Coating 
Miracool silicon acrylic emulsion paint 0.89 0.89 113 229 and 231  
Renolit Alkorbright polyester-reinforced roofing membrane 0.91 0.84 115 229 and 232  
APOC 256X coating 0.94 0.88 120 229 and 233  

Energy saving benefits from installing a cool roof can be evaluated from both modeling and field tests, which is highly dependent on local weather and climate conditions, construction of building (e.g., roof insulation levels), and the use of the building (e.g., residential or commercial buildings). A study by the Lawrence Berkeley National Laboratory (LBNL) showed that raising the roof reflectivity from an existing 10%–20% to about 60% can reduce cooling-energy use in buildings for more than 20%.234 Even more capital benefits can come from the reduced size of air conditioners.235 A few field tests have been carried out to study the impact of cool roof on building energy consumption. Pisello and Cotana236 conducted a two-year long study in Italy to investigate the effect of an innovative cool roof tile (solar reflectance: 0.77, thermal emissivity: 0.89) on residential buildings. They showed that the proposed cool roof system can reduce the peak overheating of the external roof surface by 15–18 °C in summer, which produces a maximum decrease in peak indoor overheating of the attic by up to 4.7 °C. Further analysis indicates the overall benefit corresponds to 14 kWh/m2 cooling electricity saving per year for the southern Italian climate. Testa and Krarti57 summarized both simulation studies and field tests in the literature and found that typical cooling energy savings from a cool roof with reflectance values of 0.3–0.4 for the US climates range from 0.1 to 8.6 kWh/m2 for residential buildings, 1.1–8.2 kWh/m2 for office buildings, and 1.4–10.9 kWh/m2 for retail buildings.

Though most commercially available cool roof materials can lower the surface temperature of building roofs, none of those materials can reach subambient temperatures during the day under the sun. In 2015, Gentle and Smith143 used a 67–μm-thick all-polyester material, which is a coextruded combinations of many bilayers of PET and ECDEL.237 By adding a 200-nm-thick silver coating at the bottom, the authors measured a solar reflectance of 0.97 and a thermal emissivity of 0.96 across the atmospheric window. 2 °C subambient temperature under 1060 W/m2 solar irradiance was demonstrated in a field test using a small piece of sample (19 cm × 19 cm) mounted horizontally on a polystyrene block. Another field test, conducted by the authors of this paper, compared the glass-polymer hybrid metamaterial [see Fig. 20] roof and a regular gray-colored shingle roof. Two reduced-scale model rooms have been built and tested in Laramie, Wyoming, USA, with dimensions of 2.4 m × 1.8 m × 2.4 m (width × length × height) as shown in Fig. 25(a). Temperature comparison for the roof surface and the air inside the room are shown in Figs. 25(b) and 25(c), respectively. Solar irradiation and outdoor temperature change with time are given in Fig. 25(d). It is observed that the maximum temperature difference at the roof surface and indoor were 28.6 °C and 11.2 °C, respectively. The simulation results show that the cooling load reduction from utilizing the metamaterial cool roof could save as much as 91 kWh/(m2 yr) of cooling electricity for the analyzed location of Orlando, Florida, USA, compared to the conventional shingle roof.

FIG. 25.

Comparison study of a polymer-based metamaterial cool roof and a regular shingle roof. The polymer-based metamaterial developed at the University of Colorado is shown in Fig. 20. The test was conducted at Laramie, Wyoming, USA. (a) A picture shows the two rooms for comparison study. The sizes of both rooms are 2.4 m × 1.8 m × 2.4 m (width × length × height). (b) The temperatures change of the metamaterial cool roof and shingle roof during a 24-h testing period. The maximum roof temperature difference is 28.6 °C. (c) Room temperature change during the testing period. The maximum room temperature difference is 11.2 °C. (d) Ambient temperature and solar irradiation during the testing period.

FIG. 25.

Comparison study of a polymer-based metamaterial cool roof and a regular shingle roof. The polymer-based metamaterial developed at the University of Colorado is shown in Fig. 20. The test was conducted at Laramie, Wyoming, USA. (a) A picture shows the two rooms for comparison study. The sizes of both rooms are 2.4 m × 1.8 m × 2.4 m (width × length × height). (b) The temperatures change of the metamaterial cool roof and shingle roof during a 24-h testing period. The maximum roof temperature difference is 28.6 °C. (c) Room temperature change during the testing period. The maximum room temperature difference is 11.2 °C. (d) Ambient temperature and solar irradiation during the testing period.

Close modal

Cost and payback period are important factors when considering cool roof installation. Zhang et al.238 conducted life cycle analysis of using cool roof paint in the tropical country of Singapore. Annual cooling energy savings of $33/m2 can be achieved. The payback period is expected to be less than two months in unventilated roof and less than six months in a ventilated roof. Testa and Krarti57 reviewed the benefits of cool roof and found that annual energy cost savings from cool roofs ranges from $0.16/m2 to $0.36/m2.

Even though the energy saving benefits are obvious, there are also challenges associated with cool roof applications including:

  1. The use of a cool roof could increase heating energy in winter. This problem becomes more severe for those climates where heating seasons are longer than cooling seasons. For example, at high latitudes, the use of cool roof is not as effective for reducing the energy consumption due to low solar irradiation. For a year-round operation of cool roofs, one needs to make sure that the cool roof could produce substantial benefits in summer along with relatively small or negligible penalties in winter. The cool roof technique could easily find its application at low latitudes (e.g., southeast Asia239), where cooling is a major concern for buildings. At high latitudes, one potential solution is to implement a switchable cool roof,57,207 which has the ability to change roof reflectance when buildings transition to the heating mode from the cooling mode (see Sec. III B—self-adaptive/switchable radiative sky cooling material).

  2. Another challenge of cool roof is the visual discomfort caused by such highly reflective roofs. To solve this problem, researchers have been working on increasing solar reflectivity and thermal emissivity without affecting the existing color preference, which is a compromise between roof temperature reduction and visual aesthetics.240–243 For example, the cool-colored roofing technology, which can make solar-reflective roofing available in any color (dark or light) by selectively reflecting the invisible component of sunlight in the near-infrared spectrum, which accounts for 52% of total solar power.234Figure 26 shows a few cool-colored roof tile coatings that has higher solar reflectance. In the visible spectrum between 0.4 μm and 0.7 μm, the cool-colored roof tile does not have too much difference with the standard roof, as shown in Fig. 26(a). But in the near infrared between 0.7 μm and 2.5 μm, the cool-colored roof has much higher reflectance. This result in higher solar reflectance without changing too much the color of the roof tile, as shown in Fig. 26(b).

  3. The performance of the cool roof could degrade over time due to dust accumulation and material degradation. Soiling and accumulation of soot particles can reduce solar reflectance by about 0.15 with most degradation occurring in the first year.244 Even though the solar reflectance degradation due to soiling can be effectively restored by simply washing,245 the degradation due to material itself is usually more difficult to restore. Fortunately, the thermal emissivity of the cool roof material does not degrade significantly over time.244 

FIG. 26.

Cool-colored roofing materials.234 (a) Solar spectrum reflectance of the cool and standard brown surface, as an example for demonstrating the principle of cool color roofing materials. Normalized AM 1.5 solar spectrum is shown for visualization. (b) Palette of color-matched cool (top row) and conventional (bottom row) roof tile coatings developed by American Roof Tile Coatings. Shown on each coated tile is its solar reflectance Adapted with permission from Akbari et al., Cool Color Roofing Materials. Copyright 2006 awrence Berkeley National Laboratory. Lawrence Berkeley National Laboratory.

FIG. 26.

Cool-colored roofing materials.234 (a) Solar spectrum reflectance of the cool and standard brown surface, as an example for demonstrating the principle of cool color roofing materials. Normalized AM 1.5 solar spectrum is shown for visualization. (b) Palette of color-matched cool (top row) and conventional (bottom row) roof tile coatings developed by American Roof Tile Coatings. Shown on each coated tile is its solar reflectance Adapted with permission from Akbari et al., Cool Color Roofing Materials. Copyright 2006 awrence Berkeley National Laboratory. Lawrence Berkeley National Laboratory.

Close modal

Wide use of cool roofs can benefit not only buildings, but also cities by mitigating the global warming effect. Oleson et al.246 found that the use of white roofs could decrease urban daily maximum temperature by 0.6 °C and daily minimum temperature by 0.3 °C. A few energy standards, such as ASHRAE 90.1–2016,247 ASHRAE 90.2–2007,248 and California Title 24249 have already implemented the cool roof standard. Other rating standards, such as LEED,250 also give incentive to the implementation of cool roof technology.

2. Active applications

Though cooling energy or cold harnessed through radiative sky cooling is totally free, some external energy sources (e.g., fan and pump) are needed to transfer and then to use the cooling energy. Active systems, including air-based,251–256 open or closed loop water-based,37,257–260 and hybrid systems, have been proposed to integrate nighttime radiative sky cooling with other energy systems (e.g., photovoltaic-thermal36,261). The efficiency of those systems is usually determined by radiative cooling surface properties and climate conditions (e.g., ambient air temperature, humidity, wind speed). A typical air-based radiative system consists of a radiative cooling surface (e.g., roof-mounted radiator or a cool roof), an air circulating loop, and a fan to drive the air. Lu et al.34 compared different system configurations of the nighttime radiative sky cooling system. Figure 27(a) shows a system that uses the roof as a radiator to lower the indoor temperature. Figures 27(b) and 27(c) show two schematic systems that combine the air-based system with building air conditioner for energy saving. The average net cooling power of air-based systems is usually low, between 20 and 30 W/m2.253,262 The reason could possibly be due to the low heat transfer coefficient of the air. In general, air-based systems are simpler to implement without freezing concerns. They can provide instantaneous cooling during the nighttime for buildings.252 

FIG. 27.

Building-integrated nighttime radiative sky cooling systems. (a) Air-based radiative sky cooling system that uses the roof as a radiator. (b) Air-based radiative sky cooling system integrated with a building air conditioner. (c) NightSolar® system utilize a ventilation roof design to allow both space cooling in summer and solar space heating in winter.263 (d) A radiative cooling system that takes advantage of nighttime radiative sky cooling and convection cooling. (e) Nighttime radiative sky cooling system that pumps cooled water into slab floors at night, storing the cooling energy in the building's thermal mass. (f) Nighttime radiative sky cooling for cold water production. The cold water produced can be used in various applications.37 For example, during the day, cool water stored in the tank can be pumped through interior fan-coil units to provide on-demand cooling.

FIG. 27.

Building-integrated nighttime radiative sky cooling systems. (a) Air-based radiative sky cooling system that uses the roof as a radiator. (b) Air-based radiative sky cooling system integrated with a building air conditioner. (c) NightSolar® system utilize a ventilation roof design to allow both space cooling in summer and solar space heating in winter.263 (d) A radiative cooling system that takes advantage of nighttime radiative sky cooling and convection cooling. (e) Nighttime radiative sky cooling system that pumps cooled water into slab floors at night, storing the cooling energy in the building's thermal mass. (f) Nighttime radiative sky cooling for cold water production. The cold water produced can be used in various applications.37 For example, during the day, cool water stored in the tank can be pumped through interior fan-coil units to provide on-demand cooling.

Close modal

Water-based radiative cooling systems utilize water as a heat transfer fluid. Depending on whether there is a cold storage unit (e.g., a water tank264), water-based systems can serve either as an in-situ cooling source [Figs. 27(d) and 27(e)] or as a cold storage system [Fig. 27(f)]. The average net cooling power of water-based systems are usually between 40 and 80 W/m2,36,37,265 higher than that of the air-based systems. By using a radiative cooling system that consists of a 5.3 m2 radiator and a 280-l water tank in Oslo, Norway, Meir et al.37 concluded that except for midsummer with extremely high relative humidity, radiative sky cooling can cover the cooling need of single-family houses. Flat panels that are similar to typical solar collectors can be used for cold energy collection.37,265–267 The radiative cooling panels can be integrated as a roof component.268 The cooling energy (i.e., cold water) stored due to nighttime radiative cooling can provide cold water for daytime cooling. However, since the supply temperature of chilled water from the building air conditioning system is usually low (e.g., 7 °C), water from the radiative sky cooling system might not be sufficiently chilled for direct use. Possible solutions are to integrate the water-based cooling system with the radiant floor/ceiling system that do not need a very low supply temperature of water, e.g., 15–18 °C, or use radiative-cooled water to precool the heat transfer fluid before it enters the chiller. Integrating the water-based radiative sky cooling system with other energy systems has also been explored for energy saving and efficiency improvement. For example, hybrid systems can integrate nighttime radiative sky cooling with phase-change materials for effective cold storage,25 with direct/indirect evaporative cooling units (e.g., cooling towers) to improve its efficiency,269,270 with desiccant systems to enhance cooling performance,271 with solar collectors to obtain both solar heating and nighttime radiative sky cooling,272,273 with photovoltaic-thermal panels to generate electricity and heating in the day, and cooling at night,36,261 and with ground source heat pumps to solve thermal accumulation problem in cooling dominated regions.274 

Recent advancement of daytime radiative sky cooling has rendered this technology to new possibilities in buildings.38,51,275,276 The building attic usually has much higher temperatures than the indoor environment during the day,277,278 which inevitably increase the air conditioner load due to heat transfer from the attic to the conditioned living space. Current attic ventilation directly uses ambient air. With subambient radiative cooling of air during the day, the temperature of the ventilation air into the attic can be reduced, as shown in Fig. 28(a). Another application is to precool the fresh air intake from the outdoor,279 which can reduce the cooling load of air conditioning systems, as shown in Fig. 28(b). Based on a low-cost daytime radiative sky cooling metamaterial,32 the authors of this work have developed a radiative air cooler that has a surface area of 1.84 m × 0.58 m, as shown by a picture in Fig. 28(c). Results on a typical summer day with clear sky conditions are shown in Fig. 28(d). Subambient cooling of moving air with a decent flow rate was demonstrated for the first time during the day under the sunlight. It can be observed that the radiative air cooler can cool intake outdoor air by 2–3 °C at noon under a flow rate of 7.9 l/(m2 s).

FIG. 28.

Air-based daytime radiative sky cooling for buildings. (a) Air-based radiative sky cooling system to reduce the temperature in the building attics. (b) Air-based radiative sky cooling to precool the building intake air. (c) Photo of the daytime radiative air cooler developed at the University of Colorado Boulder. (d) Test of the radiative air cooler on August 11, 2018, at Boulder, Colorado, USA. At flow rate 7.9 l/(m2 s), intake outdoor air temperature can be reduced by 2–3 °C at noon. The benefit from using such a radiative air cooler comes from two aspects: (1) reduced solar absorption by roofing materials and (2) lower inlet air temperature for attic ventilation.

FIG. 28.

Air-based daytime radiative sky cooling for buildings. (a) Air-based radiative sky cooling system to reduce the temperature in the building attics. (b) Air-based radiative sky cooling to precool the building intake air. (c) Photo of the daytime radiative air cooler developed at the University of Colorado Boulder. (d) Test of the radiative air cooler on August 11, 2018, at Boulder, Colorado, USA. At flow rate 7.9 l/(m2 s), intake outdoor air temperature can be reduced by 2–3 °C at noon. The benefit from using such a radiative air cooler comes from two aspects: (1) reduced solar absorption by roofing materials and (2) lower inlet air temperature for attic ventilation.

Close modal

Water-based daytime radiative sky cooling systems usually can produce more energy savings. Wang et al.276 introduced a nanophotonic radiative sky cooling system for office buildings, as shown in Fig. 29(a). The system has a radiative sky cooling loop and a space-cooling loop. The radiative sky cooling loop circulates water through a roof-mounted radiator whenever the temperature at the outlet of the radiator is lower than the temperature of water at the outlet of the tank. It was shown that the photonic radiative sky cooling system saved between 45% and 68% cooling electricity relative to the variable air volume (VAV) system and between 9% and 23% relative to the nighttime radiative cooling system featured with the best commercially available coating in the market. Zhang et al.38 proposed a hybrid radiative sky cooling system that stores cooling energy harnessed from the radiative cooling system in a cold storage tank, as shown in Fig. 29(b). The cooling energy is then used to precool the conditioned air to reduce the cooling load on the air conditioner for residential applications. In comparison with the electricity consumption of a split air conditioner alone, the hybrid radiative sky cooling system could save annual cooling electricity by 26% to 46% for the modeled locations (Orlando, San Diego, San Francisco, Denver).

FIG. 29.

Water-based daytime radiative sky cooling applications in buildings. (a) A photonic radiative cooling system.276 Adapted with permission from Wang et al., Renewable Energy 118, 265–277 (2018). Copyright 2017 Elsevier. (b) A hybrid radiative cooled-cold storage cooling system.38,280 Adapted with permission from Zhang et al., Appl. Energy 224 371–381 (2018), Copyright 2018 Elsevier. (c) A daytime radiative sky cooling system that is used to cool condenser to improve air-conditioner's efficiency.51 Adapted with permission from Goldstein et al., Nat. Energy 2, 17143 (2017), Copyright 2017 Macmillan Publishers Limited. (d) A radiative sky cooling system that uses both daytime and nighttime radiative sky cooling to maximize the system efficiency.52 Adapted with permission from Zhao et al., Joule 3, 111–123 (2019). Copyright 2018 Elsevier Inc.

FIG. 29.

Water-based daytime radiative sky cooling applications in buildings. (a) A photonic radiative cooling system.276 Adapted with permission from Wang et al., Renewable Energy 118, 265–277 (2018). Copyright 2017 Elsevier. (b) A hybrid radiative cooled-cold storage cooling system.38,280 Adapted with permission from Zhang et al., Appl. Energy 224 371–381 (2018), Copyright 2018 Elsevier. (c) A daytime radiative sky cooling system that is used to cool condenser to improve air-conditioner's efficiency.51 Adapted with permission from Goldstein et al., Nat. Energy 2, 17143 (2017), Copyright 2017 Macmillan Publishers Limited. (d) A radiative sky cooling system that uses both daytime and nighttime radiative sky cooling to maximize the system efficiency.52 Adapted with permission from Zhao et al., Joule 3, 111–123 (2019). Copyright 2018 Elsevier Inc.

Close modal

It is important to note that the temperature at the outlet of the radiative sky cooling panels always varies with the ambient temperature, which means that temperature of the heat transfer fluid from the cooling panels is much lower at night than during the day. Therefore, for a 24-h cycle, the lowest temperature in the cold storage tank should always occur in the early morning before sunrises. After sunrise, the ambient temperature goes up and the temperature of heat transfer fluid from radiative sky cooling panels also increases and can be higher than the cold storage tank. Therefore, the radiative cooling loop would be forced to stop. The benefit from daytime radiative sky cooling is therefore reduced without a good operation strategy.

Goldstein et al.51 proposed to use daytime radiative sky cooling based on reducing the temperature of the air conditioner's condenser to improve the overall efficiency. It is estimated that the electricity consumption can be reduced by 3%–5% for every 1 °C reduction of condenser temperature.281 They built a circulated daytime radiative water cooling panels and showed that a temperature reduction of between 3 and 5 °C under a water flow rate of 0.2 l/(min m2) can be achieved. The cooled water is then used to subcool the refrigerant from the condenser through a heat exchanger, as shown in Fig. 29(c). The modeling results show a 21% reduction in cooling electricity use for buildings. However, this approach applies to daytime use only. Considering that there could be limited or even no cooling load at night, the integration of a cold storage unit is truly beneficial to taking advantage of the cooling effect 24 h a day. Recently, Zhao et al.52 introduced a radiative cooled cold collection and storage (RadiCold) system that can provide continuous day-and-night cooling. In this system, daytime and nighttime radiative sky cooling are used separately, as depicted in Fig. 29(d). During the day the subambient temperature heat transfer fluid generated is used to directly cool the condenser, while at night the RadiCold system stores the cooling energy in a storage unit, which is retrieved during the day to reduce the cooling load on air conditioners.

Zhao et al.52 built a kW-scale RadiCold system at the University of Colorado Boulder, USA. Figure 30(a) shows the schematic of the radiative sky cooling module. Figure 30(b) shows a picture of the radiative sky cooling system that consists of 10 radiative cooling modules arranged in a 2 × 5 array. Each module has a surface area of 1.35 m2, giving a total radiative cooling surface area of 13.5 m2. Figure 30(c) shows a typical test result of three consecutive days. The inlet water temperature tracks the ambient temperature throughout the test. The net cooling power is calculated from the inlet/outlet temperature difference by using the equation q=cṁΔTA, where A is 13.5 m2. Figure 30(d) shows that the net cooling power per square meter fluctuates between 40 and 100 W/m2 during the testing period. The maximum net cooling power at night is 96 W/m2 (1296 W for the 13.5 m2 system) when the water was cooled to 3.1 °C subambient on average, and the average cooling power at noon (12–2 pm) is 45 W/m2 (607 W for the system) under an average of 952 W/m2 solar irradiance on July 1. A modeling tool based on EnergyPlus and MATLAB has been developed for evaluating the monthly electricity saving for cooling a 5000 m2 commercial office building. Modeling results show that for three different subtropical locations (Phoenix, Houston, and Miami) in the United States, the RadiCold system (810 m2 radiative cooling surface area) could potentially save 64%–82% of the electricity consumption for cooling in winter (from November to February), and save 32%–45% of the electricity consumption for cooling in summer (from May to August), as shown in Figs. 30(e) and 30(f).

FIG. 30.

Radiative sky cooling system coupled with heat transfer fluid for generating the continuous cooling effect.52 (a) Schematic of a radiative cooling device with the heat transfer fluid. (b) Picture of a radiative cooling system that has several radiative cooling panels connected in series. Water is used as the heat transfer fluid. The system has a total of 13.5 m2 radiative cooling surface area. (c) Results of a 3-day continuous test. Inlet temperature tracks ambient temperature. The test was conducted at Boulder, Colorado, USA. (d) Net cooling power change during the 3-day test. (e) Monthly electricity consumption of a commercial office building located in Phoenix, Arizona with and without the RadiCold system. (f) Monthly electricity saving by the RadiCold system at three different locations, Phoenix, Houston, and Miami. Reproduced with permission from Zhao et al., Joule 3, 111–123 (2019). Copyright 2018 Elsevier Inc.

FIG. 30.

Radiative sky cooling system coupled with heat transfer fluid for generating the continuous cooling effect.52 (a) Schematic of a radiative cooling device with the heat transfer fluid. (b) Picture of a radiative cooling system that has several radiative cooling panels connected in series. Water is used as the heat transfer fluid. The system has a total of 13.5 m2 radiative cooling surface area. (c) Results of a 3-day continuous test. Inlet temperature tracks ambient temperature. The test was conducted at Boulder, Colorado, USA. (d) Net cooling power change during the 3-day test. (e) Monthly electricity consumption of a commercial office building located in Phoenix, Arizona with and without the RadiCold system. (f) Monthly electricity saving by the RadiCold system at three different locations, Phoenix, Houston, and Miami. Reproduced with permission from Zhao et al., Joule 3, 111–123 (2019). Copyright 2018 Elsevier Inc.

Close modal

Solar cells convert the solar energy directly into electricity through the photovoltaic effect. A commercial solar cell usually has an energy conversion efficiency between 12% and 20%.282 For a solar cell material that has a specific band gap, such as 1.1 eV for Si, Fig. 31(a) shows that photons with energies below this band are not absorbed and converted to electricity, while photons with energies above the band gap are not fully converted to electrical energy.283 Therefore, a large portion of solar irradiation eventually becomes heat and heats up the solar cell. Unfortunately, solar cells are susceptible to the unwanted solar heating effect that both degrades their efficiency and shortens their lifetime.284 The operating temperature of commercially-available solar cells can easily exceed 60 °C on a sunny summer day. The relative efficiency of crystalline silicon solar cells is estimated to drop at a rate of 0.45% every 1 °C temperature increase.149 The degradation rate of the solar cell can be doubled with every 10 °C increase in the operating temperature.

FIG. 31.

Radiative sky cooling for solar cells. (a) Schematic showing that photons with energies below the band gap (Eg, denoted by the dashed black line) of the solar cell material are not absorbed and converted to electricity (left side of the dashed black line), and photons with energies above the band gap (right side of the dashed black line) are not fully converted into electrical energy with the excessive energy heating up the cell.283 Reproduced with permission from Polman et al., Science 352(6283), aad4424 (2016). Copyright 2016 American Association for the Advancement of Science. (b) Normal-view and side-view SEM images of the 2-dimensional silica photonic crystal structure. The photonic crystal consists of a square-lattice with a periodicity of 6 μm made by etching 10-μm-deep air holes into a 500-μm-thick double-side-polished fused silica wafer.40 (c) Measured emissivity/absorptivity of solar absorbers at 10° angle of incidence between the wavelengths of 6 μm and 25 μm, averaged over both polarizations. The black, blue, and red curves show the measured emissivity/absorptivity for the bare absorber structure, the absorber structure with the planar silica layer, and the absorber structure with the silica photonic crystal, respectively. (d) Steady-state temperature measurement showing that the silica photonic crystal is on average 1.3 °C cooler than the absorber structure with the planar silica layer. (b)–(d) are reproduced with permission from Zhu et al., PNAS 112(40), 12282–12287 (2015). Copyright 2015 National Academy of Sciences.

FIG. 31.

Radiative sky cooling for solar cells. (a) Schematic showing that photons with energies below the band gap (Eg, denoted by the dashed black line) of the solar cell material are not absorbed and converted to electricity (left side of the dashed black line), and photons with energies above the band gap (right side of the dashed black line) are not fully converted into electrical energy with the excessive energy heating up the cell.283 Reproduced with permission from Polman et al., Science 352(6283), aad4424 (2016). Copyright 2016 American Association for the Advancement of Science. (b) Normal-view and side-view SEM images of the 2-dimensional silica photonic crystal structure. The photonic crystal consists of a square-lattice with a periodicity of 6 μm made by etching 10-μm-deep air holes into a 500-μm-thick double-side-polished fused silica wafer.40 (c) Measured emissivity/absorptivity of solar absorbers at 10° angle of incidence between the wavelengths of 6 μm and 25 μm, averaged over both polarizations. The black, blue, and red curves show the measured emissivity/absorptivity for the bare absorber structure, the absorber structure with the planar silica layer, and the absorber structure with the silica photonic crystal, respectively. (d) Steady-state temperature measurement showing that the silica photonic crystal is on average 1.3 °C cooler than the absorber structure with the planar silica layer. (b)–(d) are reproduced with permission from Zhu et al., PNAS 112(40), 12282–12287 (2015). Copyright 2015 National Academy of Sciences.

Close modal

Earlier works on cooling technologies for solar cells include surface water cooling,285,286 natural or forced air ventilation,42,287 heat pipes,288 thermoelectric modules,289 and phase change material.290,291 Radiative sky cooling of solar cells could be a straightforward idea and can be made possible by placing a thin-layer material that has high transmittance in the solar spectrum and high emissivity over the infrared spectrum at the operating temperature. Different from building applications that seek to achieving subambient temperatures, the major objective of solar cell cooling is to lower its operating temperature as much as possible while maintaining its solar absorptance,292 although Safi and Munday293 theoretically showed that solar cells can be cooled to subambient temperature with radiative sky cooling. Therefore, a wavelength-selective radiative emitter at the atmospheric window is not absolutely necessary. A broadband surface is instead preferred for solar cell cooling, especially when considering the cost benefits.294 

Standard solar cells are typically covered with glass, which has a mid-infrared emissivity of 0.84. Riverola et al.295 deduced that the mid-infrared emissivity of a PV module is very high and almost independent of the underlying solar cell structure. Subedi et al.296 calculated emissivity from specular and diffuse reflectance measurements of three commercial low iron soda lime glasses commonly used in PV modules: Solite™, Diamant®, and Pilkington. The authors concluded that all three glasses have almost the same hemispherical emissivities and are greater than 0.84. However, Gentle and Smith294 suggested that the emissivity value of glass (i.e., 0.84) could have been overestimated and suggested to hemispherical emissivity to be around 0.75. The higher the emissivity of glass cover, the more challenging to obtain the additional radiative sky cooling effect by enhancing mid-infrared thermal radiation. However, the thermal emissivity of typical glass is not optimized in terms of both wavelength and incident angle, which suggests that there still remain some rooms to enhance the radiative cooling power. Zhu et al.292 showed theoretically that an ideal radiative emitter (emissivity equals to 1) on a solar cell could operate at a temperature 5.2 K below that with a 5-mm-thick fused silica glass (emissivity around 0.73). Then the same group of authors proposed a structure using silica nanophotonic crystal [Fig. 31(b)] that has near-unity emissivity across the mid-infrared wavelength range [Fig. 31(c)] on top of the silicon absorber.40 The field temperature test showed that the silica absorber with the nanophotonic crystal structure is >1 °C cooler than with the planar silica layer at noon, as shown in Fig. 31(d). Even though this is not a big temperature reduction, the efficiency improvement is still worth pursuing if no extra infrastructure is needed and the nanophotonic structure can be manufactured in a cost-effective way.294 

Radiative sky cooling is more appealing for concentrated photovoltaic (CPV) systems which usually operates at even higher temperatures.79 By using a 30–μm-thick radiative cooling layer consists of acrylate resin and inorganic fillers (coated on the aluminum chassis of a CPV module), Nishioka et al.297 experimentally demonstrated 10 °C temperature reduction (from 93.1 °C to 82.6 °C) of a CPV module compared to that without the radiative cooling layer. The employment of the radiative cooling coating also showed considerably improved uniformity of the temperature distribution in the CPV module. The open-circuit voltage of the CPV module with radiative cooling coating was 0.5 V higher than that of the module without the coating during the test period, which gives 0.5% higher conversion efficiency.

Published studies on radiative sky cooling for solar cells are limited to theoretical analysis,41,292 with a few field tests showing solar cell temperature reduction compared to that with no radiative sky cooling.40,297,298 There still lacks experimental investigation on the photocurrent-voltage (I-V) characteristics of solar cells.205 The other challenges are long-term durability and cost-effectiveness. Although some researchers claimed that their materials can be made on a large-scale and in a cost-effectively way,299 it is still unclear whether radiative sky cooling can be both technically feasible and economically viable.

Recently a combined approach to cool solar cells by simultaneously reflecting of sub-band-gap and ultraviolet solar irradiation, and enhancing radiative sky cooling was proposed, as shown in Fig. 32(a). Li et al.41 proposed a multilayer dielectric stack coating on top of an existing encapsulated solar panel, which was designed to simultaneously reflect sub-band-gap solar irradiation and to enhance radiative sky cooling, as shown in Fig. 32(b). Based on calculated solar reflectivity [Fig. 32(c)] and thermal emissivity [Fig. 32(d)] spectra, the photonic coating structure can provide a temperature reduction of ∼5.7 K as compared to the same solar cell without the coating. Another simulation study by Sun et al.300 shows that the combined technique would cool one-sun terrestrial solar modules up to 10 °C. This combined technique is very promising for reducing the solar cell temperature while further research should be focused on field testing.301 

FIG. 32.

Cooling of solar cells by both radiative sky cooling and sub-band-gap reflection.41 (a) A schematic showing that radiative sky cooling can be enhanced and the sub-band-gap solar radiation can be strongly reflected by using a nanophotonic cooler coated on top of an existing encapsulated solar panel. (b) A schematic of the nanophotonic cooler made of a multilayer dielectric stack that consists of alternating layers of Al2O3/SiN/TiO2/SiN with aperiodic arrangement of thickness. (c) Reflection spectrum of the photonic cooler (blue) with n = 11 sublayers. Near unity reflectivity is achieved in the wavelength range between 1.3 and 1.8 μm. (d) Emissivity spectra of the photonic cooler show broadband high emissivity over the infrared wavelength (4–25 μm). Reproduced with permission from Li et al., ACS Photonics 4, 774–782 (2017). Copyright 2017 American Chemical Society.

FIG. 32.

Cooling of solar cells by both radiative sky cooling and sub-band-gap reflection.41 (a) A schematic showing that radiative sky cooling can be enhanced and the sub-band-gap solar radiation can be strongly reflected by using a nanophotonic cooler coated on top of an existing encapsulated solar panel. (b) A schematic of the nanophotonic cooler made of a multilayer dielectric stack that consists of alternating layers of Al2O3/SiN/TiO2/SiN with aperiodic arrangement of thickness. (c) Reflection spectrum of the photonic cooler (blue) with n = 11 sublayers. Near unity reflectivity is achieved in the wavelength range between 1.3 and 1.8 μm. (d) Emissivity spectra of the photonic cooler show broadband high emissivity over the infrared wavelength (4–25 μm). Reproduced with permission from Li et al., ACS Photonics 4, 774–782 (2017). Copyright 2017 American Chemical Society.

Close modal

Dew forms on a surface when the surface temperature is below the local dew point temperature. Dew harvesting could generate a significant impact on water supply in certain climates/regions,302,303 for example, in many arid and semi-arid areas where traditional water sources are subject to depletion. The atmosphere air is a huge and renewable reservoir that contains 12 900 cubic kilometers of water.304 The availability of atmosphere water has no geographic dependence. Moreover, dew water is generally potable, especially in rural and isolated locations (e.g., islands), where it is most beneficial.47,305,306

There are generally three ways to harvest water from air: adsorption technology based on using adsorbent,307–309 surface cooling by using heat pumps,310 and radiative sky cooling. Research works on radiative sky cooling dew condensers can be traced back to the 1960s.311 Efforts have been devoted to analyze climate and weather conditions,312,313 to develop radiative cooling surface materials,171,314 to design dew collectors,169 and to study operating conditions that affect the water production rate.46,315–317 The dilemma for the radiative sky cooling dew harvester is that the local atmosphere should have a large amount of water vapor (i.e., high humidity) that can be condensed, while the presence of a large amount of water vapor inevitably limits the radiative cooling power due to the infrared absorption of water. While the surface should allow easy condensate nucleation, the condensate needs to be removed promptly right after condensation happens on the surface.

The theoretical limit of the dew yield is about 0.8 l/(m2·day) which is restricted by the available cooling power (∼100 W/m2) with respect to the latent heat of water condensation (2260 kJ/kg).313 Nevertheless, in practice the dew yield is much less than this limit due to the imperfect climate condition and system design. Table IV lists the performance of some dew water harvesting devices using nighttime radiative sky cooling. Atmospheric emissivity, relative humidity, and local wind speed all affect dew water condensation. It was found that dew yields dropped linearly with the increase in the cloud coverage.313 However, a completely clear sky does not necessarily give highest dew yield because a clearer sky is usually associated with drier air, indicating that a certain level of absolute humidity is necessary for dew condensation to occur.313 The relative humidity has to be high to reduce the required temperature difference (between ambient and dew point temperatures) for condensation to occur. While a high wind speed is not favorable for dew formation because it reduces the net cooling power, and a gentle wind speed (less than 1 m/s) is necessary for bringing atmospheric water vapor to the condenser (i.e., radiative cooling surface).306,320 Researchers have been exploring the correlations between meteorological data (relative humidity, cloud cover, and wind speed) with dew water yields to predict the water production rate under different climate conditions.

TABLE IV.

Dew water harvesting enabled by nighttime radiative sky cooling.

MaterialLocation (climate condition)Cooling Performance (condensation rate)Reference
Thin foil (0.39 mm thick) made of polyethylene embedded with TiO2 and BaSO4 microspheres, and a polyvinyl chloride plastic sheet (4 mm) Netherlands (marine climate) Average 0.15 ± 0.05 l/(m2 day) over a 1.5-year period Jacobs et al.169  
Thin foil (0.39 mm thick) made of small TiO2 and BaSO4 microspheres embedded in a matrix of low-density polyethylene (LDPE) Croatia (Mediterranean coast) Yearly average 0.055 l/(m2 day) in Zadar and 0.025 l/(m2 day) in Komiza Muselli et al.313  
Pigmented polyethylene foils (pigments are TiO2 and BaSO4 microspheres embedded in a 350 μm thickness polyethylene foil Dhahran, Saudi Arabia (dry) 0.22 l/(m2 day) during a single night of operation Gandhidasan and Abualhamayel318  
Galvanized iron roof Kothara, India (hot and dry) Average 0.098 l/(m2 day) during the dry season Sharan et al.44  
Pigmented polymer foils (pigments are the SiO2/TiO2 composite) Dodoma, Tanzania (semi-arid) Average 0.04 l/(m2 day) during a dry month Nilsson et al.319  
MaterialLocation (climate condition)Cooling Performance (condensation rate)Reference
Thin foil (0.39 mm thick) made of polyethylene embedded with TiO2 and BaSO4 microspheres, and a polyvinyl chloride plastic sheet (4 mm) Netherlands (marine climate) Average 0.15 ± 0.05 l/(m2 day) over a 1.5-year period Jacobs et al.169  
Thin foil (0.39 mm thick) made of small TiO2 and BaSO4 microspheres embedded in a matrix of low-density polyethylene (LDPE) Croatia (Mediterranean coast) Yearly average 0.055 l/(m2 day) in Zadar and 0.025 l/(m2 day) in Komiza Muselli et al.313  
Pigmented polyethylene foils (pigments are TiO2 and BaSO4 microspheres embedded in a 350 μm thickness polyethylene foil Dhahran, Saudi Arabia (dry) 0.22 l/(m2 day) during a single night of operation Gandhidasan and Abualhamayel318  
Galvanized iron roof Kothara, India (hot and dry) Average 0.098 l/(m2 day) during the dry season Sharan et al.44  
Pigmented polymer foils (pigments are the SiO2/TiO2 composite) Dodoma, Tanzania (semi-arid) Average 0.04 l/(m2 day) during a dry month Nilsson et al.319  

Worldwide power production is rapidly increasing to meet the expanding needs of growing world population and technological advancements. Currently, water-cooled thermal power plants are the mainstream because of their high thermal efficiency, in which the waste heat is generally being removed by cooling towers and cooling ponds. In the US alone, more than 99% of thermal power plants utilize some form of water cooling technologies today,50 and power plants use 41% of the freshwater draw with 3% of water lost to the air, which is a direct challenge to resolving widespread water scarcity and frequent droughts.321 Globally, water scarcity is a serious issue, especially in those areas such as northern Africa, the Middle East, and India where electrification is also projected to grow rapidly. Air-cooled (i.e., dry-cooled) techniques, as an alternative to water cooling, trade the water consumption with plant efficiency, cost (accounting for up to a 9% increase in the levelized cost of electricity [LCOE]), and robustness (power plant efficiency derates 1% per 3 °C increase in ambient temperature above the design point), as shown in Fig. 33(a). In such circumstances, passive and enhanced cooling technologies are appealing. Recent studies propose that radiative sky cooling can be a supplemental cooling option for both wet-cooled and dry-cooled thermal power plants.150,322 Radiative sky cooling could potentially both reduce the amount of water to be evaporated by cooling towers and out-perform air-cooled systems in terms of efficiency.

FIG. 33.

Radiative sky cooling for power plant applications. (a) Variation in power plant energy conversion efficiency and levelized cost of electricity (LCOE) as a function of ambient air temperature for the air cooled heat exchanger. (b) Energy density mismatch is the biggest obstacle for implementing radiative sky cooling in power plants. A two-stage radiative cooled-cold collection and storage system was proposed to overcome the energy density mismatch problem for power plant supplementary cooling.52,325 The first stage is radiative cooling modules for local cold collection and storage. The selective surface structure is to be laminated on top of the module to provide cooling power. The second stage storage is a larger size cold water storage tank, which connects to multiple radiative cooling modules. (c) Power production and heat rejection by the power block and cold generation capacity of the radiative cooling system. ΔTapproach is the difference between the condenser inlet temperature and the web-bulb ambient temperature. The radiative sky cooling system is assumed to have the same size as and the CSP power plant solar field. The heat rejection by the power block and cold generation by the cooling system both increase, with the latter at a higher rate. (d) Annual cooling water make-up as a function of ΔTapproach. Without the radiative sky cooling system, the cooling water make-up increases due to the increasing heat rejection. With the radiative sky cooling system, the cooling water make-up decreases because the system cold generation capacity goes up at a higher rate than the heat rejection.

FIG. 33.

Radiative sky cooling for power plant applications. (a) Variation in power plant energy conversion efficiency and levelized cost of electricity (LCOE) as a function of ambient air temperature for the air cooled heat exchanger. (b) Energy density mismatch is the biggest obstacle for implementing radiative sky cooling in power plants. A two-stage radiative cooled-cold collection and storage system was proposed to overcome the energy density mismatch problem for power plant supplementary cooling.52,325 The first stage is radiative cooling modules for local cold collection and storage. The selective surface structure is to be laminated on top of the module to provide cooling power. The second stage storage is a larger size cold water storage tank, which connects to multiple radiative cooling modules. (c) Power production and heat rejection by the power block and cold generation capacity of the radiative cooling system. ΔTapproach is the difference between the condenser inlet temperature and the web-bulb ambient temperature. The radiative sky cooling system is assumed to have the same size as and the CSP power plant solar field. The heat rejection by the power block and cold generation by the cooling system both increase, with the latter at a higher rate. (d) Annual cooling water make-up as a function of ΔTapproach. Without the radiative sky cooling system, the cooling water make-up increases due to the increasing heat rejection. With the radiative sky cooling system, the cooling water make-up decreases because the system cold generation capacity goes up at a higher rate than the heat rejection.

Close modal

Though a great potential has been identified for power plant cooling, there is not much investigation on radiative sky cooling for power plant application yet. Olwi and Sabbagh323,324 presented a modeling study to show the potential of using covered ponds as a cooling system for power plants in arid areas. Numerical study shows that a total heat rejection of 150 W/m2 could be obtained on average. The biggest challenge in utilizing radiative sky cooling for power plants is its intrinsically low energy density (∼100 W/m2), while the need for the cooling of the power plant condenser is on the order of 100 MW. This indicates the need of large area of land and high operational cost to circulate heat transfer fluid between the power plant condenser and radiative sky cooling panels. It is important to note that not all thermal power plants favor radiative sky cooling technology, such as those power plants located at water abundant areas. On the other hand, most of the concentrating solar power (CSP) plants are located in deserts where solar resource is abundant, and water is scarce. Coupling of radiative sky cooling with CSP seems a perfect match since radiative sky cooling prefers the low nighttime temperatures and clear skies.150 Assuming the nighttime radiative sky cooling system has the same size as the solar collector field, Dyreson and Miller150 performed simulations and showed that the system can provide over 90% of the required cooling. However, their feasibility study requires two massive storage tanks, and the working temperature of the water seems to be rather high (39–63 °C), which implies higher backpressure in the condenser and thus lower thermal efficiency. Zeyghami and Khalili322 investigated daytime radiative sky cooling for an air-cooled CSP plant that uses supercritical carbon dioxide (CO2) cycles. The results show that it is possible to overcome the efficiency loss due to air cooling by 3%–7.5% depending on the hot source temperature.

Zhang et al.325 recently introduced a two-stage radiative cooled-cold collection and storage system to overcome the energy density mismatch problem, as show in Fig. 33(b). The first stage is radiative cooling modules for local cold collection and storage. The selective radiative cooling surface material is laminated on top of the module. The second stage is larger size cold water storage tanks, which connect to multiple radiative cooling modules. A tree-like pipe network was proposed to minimize the pumping energy input for cold collection. Modeling has been conducted to study the benefits after applying the radiative cooling system to a wet-cooled CSP power plant (Mojave solar project). With the field aperture of the radiative cooling system the same as the power plant's solar field, up to 83% of water use can be saved if the cooling system works during both daytime and nighttime [see Figs. 33(c) and 33(d)]. Note that for some CSP systems, for example, the parabolic trough system, there already exist a solar farm and a hydraulic system for transferring thermal energy. If the solar farm can be turned into a radiative sky cooling farm at night using dual functional panels, the initial and maintenance costs for radiative sky cooling power plant application can be greatly reduced. Future research studies are still needed to identify optimal designs and operation strategies to show water savings while improving (or at least not affecting) the power plant thermal efficiencies.

With daytime subambient radiative sky cooling now demonstrated, many other possible applications can be explored, especially with a scalable-manufactured low-cost material. For example, radiative sky cooling can be used for the storage of some fruits, vegetables, or medicine in the decentralized community in off-grid remote areas or in developing countries. Ezekew326 developed a cooling box (0.6 × 0.3 × 0.5 m) with a heat pipe assisted sky radiative cooler. During a 24-h experiment, the lowest temperature in the box was 12.8 °C (ambient 20 °C) and 15 °C (ambient 32 °C) at night and during the day, respectively. The radiative cooling box's passive mode of operation and its portable nature make it suitable for use in remote areas. Other researchers have attempted to use radiative sky cooling to enhance the performance of solar still. Haddad et al.327 proposed to store the nighttime radiative sky cooling energy in the solar still's packed bed condenser, then the cooling energy can be used during the day to condense vapor produced by the solar still. Recently, Mu et al.328 demonstrated an ultrathin thermoelectric generator that can directly convert heat from the environment into electricity by using radiative sky cooling. The idea is that electricity can be generated with zero energy input. With about 4 K temperature drop (across the thermoelectric generator) achieved by using the radiative cooling emitter, the output voltage of the thermoelectric generator reached up to 0.5 mV, and exhibited a continuous average 0.18 mV for 24 h operation.

In recent years, there have been significant advances in human body cooling by using visible-opaque, but infrared-transparent textiles in the indoor environment.329,330 There is also attempt to use radiative sky cooling for human body cooling in the outdoors, which requires the textile material to have both high reflection for solar irradiation and high transmission for human body thermal radiation. Cai et al.49 developed a novel textile material by embedding zinc oxide nanoparticles (NPs) into nanoporous polyethylene (ZnO-PE), as shown in Figs. 34(a) and 34(b). The reflection of solar irradiation is more than 90%, while the transmissivity of human body thermal radiation can be higher than 80% [Fig. 34(c)]. By covering the textile material on a simulated skin, the authors showed that the material can enable the simulated skin to avoid overheating by more than 10 °C, compared to cotton under typical outdoor environment with peak solar irradiation over 900 W/m2, which has demonstrated radiative sky cooling for outdoor personal cooling. The nanoporous polyethylene can be potentially fabricated in large-scale for industrial fabric production.331 

FIG. 34.

Spectrally selective nanocomposite textile uses daytime radiative sky cooling for outdoor personal cooling.49 (a) A novel textile material that has high reflection for solar irradiation and high transmission for human body thermal radiation enabled by embedding zinc oxide nanoparticles (NPs) into nanoporous polyethylene (ZnO-PE). (b) Photo of the ZnO-PE textile material under the sun. (c) Measured (solid curves) and simulated (dashed curves) reflectivity and transmissivity spectra of the ZnO-PE textile at the wavelength range between 0.3 μm and 16 μm. Figures reproduced with permission from Cai et al. Adv. Mater. 30, 1802512 (2018). Copyright 2018 Wiley.

FIG. 34.

Spectrally selective nanocomposite textile uses daytime radiative sky cooling for outdoor personal cooling.49 (a) A novel textile material that has high reflection for solar irradiation and high transmission for human body thermal radiation enabled by embedding zinc oxide nanoparticles (NPs) into nanoporous polyethylene (ZnO-PE). (b) Photo of the ZnO-PE textile material under the sun. (c) Measured (solid curves) and simulated (dashed curves) reflectivity and transmissivity spectra of the ZnO-PE textile at the wavelength range between 0.3 μm and 16 μm. Figures reproduced with permission from Cai et al. Adv. Mater. 30, 1802512 (2018). Copyright 2018 Wiley.

Close modal

As the ultimate heat sink for the earth, the great potential of using radiative sky cooling without consuming external energy has been long recognized, but the practical applications have been largely fallen behind its potential for several decades. Recent advances on daytime radiative sky cooling materials have revived the interest world-wide in this field. In order to have the net cooling effect under the sunlight, the material should possess larger than 95% reflection of solar irradiation. As evidenced in this review, a list of nanophotonic materials,31 metamaterials,32 paints, and coatings58 have been developed. Great progresses have been made in terms of material fabrication. A scalable-manufactured polymer-based daytime radiative cooling material has been successfully demonstrated, which suggest that this technique can be applied on a large-scale at a low cost.32 The integration of daytime radiative sky cooling materials with new functionality, e.g., self-adaptive cooling,207 could play a significant role in improving the efficiency of this technique.

Great progresses have also been made in terms of demonstrating radiative sky cooling in real-world applications. The benefits of radiative sky cooling come from various aspects, including cooling electricity saving, downsizing the HVAC system in buildings, increase in solar cell power generation and efficiency gain, and water saving for power plants. For system integration, passive radiative sky cooling is currently the most promising approach due to its system simplicity, low cost, and low maintenance. Solar cell cooling by simultaneously reflecting the sub-band-gap and ultraviolet solar irradiation, and enhancing radiative sky cooling with a top layer on the existing solar cell structure is very appealing.41 For active radiative sky cooling systems, it is highly recommended to take advantage of the existing infrastructure, for example, the existing hydraulic loop and air ducts in the buildings. With daytime radiative sky cooling demonstrated, a radiative sky cooling system can be built to run 24 h continuously, which renders even larger energy savings from the system.52 

Performance of radiative sky cooling systems largely depends on the local meteorological conditions, such as the atmospheric constituents (primarily water vapor), sky conditions (i.e., clear or overcast), local wind speeds, and the constantly changing weather conditions (i.e., unstable). The effect of atmospheric water vapor on the radiative sky cooling system efficiency can be detrimental. Though radiative sky cooling effect could still be available when the sky is partially or even completely covered (if the cloud base is high), the cooling power could be substantially reduced. As such, it is important to identify those locations (climates) that are most suitable for the deployment of radiative sky cooling technology. As the energy situation and environmental issues become more severe in the 21st century, radiative sky cooling is expected to play an important role in reducing energy consumption in buildings, mitigating the urban heat island effect, resolving water and environmental issues, and even fighting against the global warming effect332 in the near future.

The authors acknowledge the financial support of this work from the U.S. Department of Energy's Advanced Research Projects Agency—Energy (ARPA-E) under Contract No. DE-AR0000580.

1.
Cooling a Warming Planet
: A Global Air Conditioning Surge, http://e360.yale.edu/features/cooling_a_warming_planet_a_global_air_conditioning_surge for Yale E360; accessed 11 November 2017.
2.
J.
Steven Brown
and
P. A.
Domanski
, “
Review of alternative cooling technologies
,”
Appl. Therm. Eng.
64
,
252
262
(
2014
).
3.
B. I.
Cook
,
J. E.
Smerdon
,
R.
Seager
, and
S.
Coats
, “
Global warming and 21st century drying
,”
Clim. Dyn.
43
,
2607
2627
(
2014
).
4.
J.
Henley
, “
World set to use more energy for cooling than heating
,” in The Guardian (
2015
).
5.
D. J.
Fixsen
, “
The temperature of the cosmic microwave background
,”
Astrophys. J.
707
,
916
(
2009
).
6.
G. B.
Smith
and
C.-G. S.
Granqvist
,
Green Nanotechnology: Solutions for Sustainability and Energy in the Built Environment
(
CRC Press
,
2010
).
7.
K. Y.
Kondrat'yev
,
Radiative Heat Exchange Atmosphere
(
Pergamon
,
1965
).
8.
S.
Chu
,
Y.
Cui
, and
N.
Liu
, “
The path towards sustainable energy
,”
Nat. Mater.
16
,
16
22
(
2017
).
9.
M. M.
Hossain
and
M.
Gu
, “
Radiative cooling: Principles, progress, and potentials
,”
Adv. Sci.
3
,
1500360
(
2016
).
10.
J. L.
Miller
, “
Passive cooling doesn't cost the planet
,”
Phys. Today
70
(
4
),
16
18
(
2017
).
11.
A. K.
Stark
, “
Methods for rejecting daytime waste heat to outer space
,”
Natl. Sci. Rev.
4
,
789
790
(
2017
).
12.
S.
Buddhiraju
,
P.
Santhanam
, and
S.
Fan
, “
Thermodynamic limits of energy harvesting from outgoing thermal radiation
,”
Proc. Natl. Acad. Sci.
115
,
E3609
E3615
(
2018
).
13.
A. K.
Head
, “
Method and means for producing refrigeration by selective radiation
,” U.S. patent US3043112A (10 July 1962).
14.
M. N.
Bahadori
, “
Passive cooling systems in Iranian architecture
,”
Sci. Am.
238
,
144
155
(
1978
).
15.
E.
Clark
and
P.
Berdahl
,
Radiative Cooling: Resource and Applications
(
Sponsor Development US Department of Energy
,
1981
), p.
219
.
16.
C. G.
Granqvist
and
A.
Hjortsberg
, “
Surfaces for radiative cooling: Silicon monoxide films on aluminum
,”
Appl. Phys. Lett.
36
,
139
141
(
1980
).
17.
P.
Berdahl
, “
Radiative cooling with MgO and/or LiF layers
,”
Appl. Opt.
23
,
370
(
1984
).
18.
C. D.
Reid
and
E. D.
McAlister
, “
Measurement of spectral emissivity from 2 μ to 15 μ
,”
JOSA
49
,
78
82
(
1959
).
19.
A.
Ben-Shalom
 et al., “
Sky radiance at wavelengths between 7 and 14 μm: Measurement, calculation, and comparison with lowtran-4 predictions
,”
Appl. Opt.
19
,
838
(
1980
).
20.
E. E.
Bell
,
L.
Eisner
,
J.
Young
, and
R. A.
Oetjen
, “
Spectral radiance of sky and terrain at wavelengths between 1 and 20 microns II sky measurements
,”
J. Opt. Soc. Am.
50
,
1313
(
1960
).
21.
M.
Hori
 et al., “
In-situ measured spectral directional emissivity of snow and ice in the 8–14 μm atmospheric window
,”
Remote Sens. Environ.
100
,
486
502
(
2006
).
22.
P.
Berdahl
and
R.
Fromberg
, “
The thermal radiance of clear skies
,”
Sol. Energy
29
,
299
314
(
1982
).
23.
M.
Martin
and
P.
Berdahl
, “
Summary of results from the spectral and angular sky radiation measurement program
,”
Sol. Energy
33
,
241
252
(
1984
).
24.
A. P.
Tsoy
,
A. S.
Granovskiy
,
A. V.
Baranenko
, and
D. A.
Tsoy
, “
Effectiveness of a night radiative cooling system in different geographical latitudes
,”
AIP Conf. Proc.
1876
,
020060
(
2017
).
25.
S.
Zhang
and
J.
Niu
, “
Cooling performance of nocturnal radiative cooling combined with microencapsulated phase change material (MPCM) slurry storage
,”
Energy Build.
54
,
122
130
(
2012
).
26.
A.
Argiriou
,
M.
Santamouris
, and
D. N.
Assimakopoulos
, “
Assessment of the radiative cooling potential of a collector using hourly weather data
,”
Energy
19
,
879
888
(
1994
).
27.
M.
Martin
and
P.
Berdahl
, “
Characteristics of infrared sky radiation in the United States
,”
Sol. Energy
33
,
321
336
(
1984
).
28.
T. S.
Eriksson
,
E. M.
Lushiku
, and
C. G.
Granqvist
, “
Materials for radiative cooling to low temperature
,”
Sol. Energy Mater.
11
,
149
161
(
1984
).
29.
R.
Family
and
M. P.
Mengüç
, “
Materials for radiative cooling: A review
,”
Procedia Environ. Sci.
38
,
752
759
(
2017
).
30.
C. G.
Granqvist
, “
Solar energy materials for thermal applications
,” in
Kirk-Othmer Encyclopedia of Chemical Technology
(
John Wiley & Sons, Inc.
,
2017
).
31.
A. P.
Raman
,
M. A.
Anoma
,
L.
Zhu
,
E.
Rephaeli
, and
S.
Fan
, “
Passive radiative cooling below ambient air temperature under direct sunlight
,”
Nature
515
,
540
544
(
2014
).
32.
Y.
Zhai
 et al., “
Scalable-manufactured randomized glass-polymer hybrid metamaterial for daytime radiative cooling
,”
Science
355
,
1062
1066
(
2017
).
33.
S.
Fan
and
A.
Raman
, “
Metamaterials for radiative sky cooling
,”
Natl. Sci. Rev.
5
,
132
133
(
2018
).
34.
X.
Lu
,
P.
Xu
,
H.
Wang
,
T.
Yang
, and
J.
Hou
, “
Cooling potential and applications prospects of passive radiative cooling in buildings: The current state-of-the-art
,”
Renewable Sustainable Energy Rev.
65
,
1079
1097
(
2016
).
35.
A.
Royne
,
C. J.
Dey
, and
D. R.
Mills
, “
Cooling of photovoltaic cells under concentrated illumination: A critical review
,”
Sol. Energy Mater. Sol. Cells
86
,
451
483
(
2005
).
36.
U.
Eicker
and
A.
Dalibard
, “
Photovoltaic–thermal collectors for night radiative cooling of buildings
,”
Sol. Energy
85
,
1322
1335
(
2011
).
37.
M. G.
Meir
,
J. B.
Rekstad
, and
O. M.
LØvvik
, “
A study of a polymer-based radiative cooling system
,”
Sol. Energy
73
,
403
417
(
2002
).
38.
K.
Zhang
,
D.
Zhao
,
X.
Yin
,
R.
Yang
, and
G.
Tan
, “
Energy saving and economic analysis of a new hybrid radiative cooling system for single-family houses in the USA
,”
Appl. Energy
224
,
371
381
(
2018
).
39.
J.
Yellott
, “
Utilization of sun and sky radiation for heating and cooling of buildings
,”
ASHRAE J. Am. Soc. Heat. Refrig. Air Cond. Eng.
15
,
31
42
(
1973
).
40.
L.
Zhu
,
A. P.
Raman
, and
S.
Fan
, “
Radiative cooling of solar absorbers using a visibly transparent photonic crystal thermal blackbody
,”
Proc. Natl. Acad. Sci.
112
,
12282
12287
(
2015
).
41.
W.
Li
,
Y.
Shi
,
K.
Chen
,
L.
Zhu
, and
S.
Fan
, “
A comprehensive photonic approach for solar cell cooling
,”
ACS Photonics
4
,
774
782
(
2017
).
42.
D.
Du
,
J.
Darkwa
, and
G.
Kokogiannakis
, “
Thermal management systems for photovoltaics (PV) installations: A critical review
,”
Sol. Energy
97
,
238
254
(
2013
).
43.
S.-H.
Wu
and
M. L.
Povinelli
, “
Solar heating of GaAs nanowire solar cells
,”
Opt. Express
23
,
A1363
A1372
(
2015
).
44.
G.
Sharan
,
D.
Beysens
, and
I.
Milimouk-Melnytchouk
, “
A study of dew water yields on Galvanized iron roofs in Kothara (North-West India)
,”
J. Arid Environ.
69
,
259
269
(
2007
).
45.
O.
Clus
 et al., “
Comparison of various radiation-cooled dew condensers using computational fluid dynamics
,”
Desalination
249
,
707
712
(
2009
).
46.
J. F.
Maestre-Valero
,
R.
Ragab
,
V.
Martínez-Alvarez
, and
A.
Baille
, “
Estimation of dew yield from radiative condensers by means of an energy balance model
,”
J. Hydrol.
460–461
,
103
109
(
2012
).
47.
G.
Sharan
, “
Harvesting dew with radiation cooled condensers to supplement drinking water supply in semi-arid
,”
Int. J. Serv. Learn. Eng. Humanit. Eng. Soc. Entrepreneur
6
,
130
150
(
2011
).
48.
D.
Carvajal
 et al., “
Roof-integrated dew water harvesting in Combarbalá, Chile
,”
J. Water Supply Res. Technol. - Aqua
67
,
357
374
(
2018
).
49.
L.
Cai
 et al., “
Spectrally selective nanocomposite textile for outdoor personal cooling
,”
Adv. Mater.
30
,
1802152
(
2018
).
50.
A. K.
Stark
and
J. F.
Klausner
, “
An R&D strategy to decouple energy from water
,”
Joule
1
,
416
420
(
2017
).
51.
E. A.
Goldstein
,
A. P.
Raman
, and
S.
Fan
, “
Sub-ambient non-evaporative fluid cooling with the sky
,”
Nat. Energy
2
,
17143
(
2017
).
52.
D.
Zhao
 et al., “
Subambient cooling of water: Towards real-world applications of daytime radiative cooling
,”
Joule
3
,
111
123
(
2019
).
53.
E.
du Marchie van Voorthuysen
and
R.
Roes
Blue sky cooling for parabolic trough plants
,”
Energy Procedia
49
,
71
79
(
2014
).
54.
C. G.
Granqvist
and
A.
Hjortsberg
, “
Radiative cooling to low temperatures: General considerations and application to selectively emitting SiO films
,”
J. Appl. Phys.
52
,
4205
4220
(
1981
).
55.
S.
Catalanotti
 et al., “
The radiative cooling of selective surfaces
,”
Sol. Energy
17
,
83
89
(
1975
).
56.
M.
Hanif
,
T. M. I.
Mahlia
,
A.
Zare
,
T. J.
Saksahdan
, and
H. S. C.
Metselaar
, “
Potential energy savings by radiative cooling system for a building in tropical climate
,”
Renewable Sustainable Energy Rev.
32
,
642
650
(
2014
).
57.
J.
Testa
and
M.
Krarti
, “
A review of benefits and limitations of static and switchable cool roof systems
,”
Renewable Sustainable Energy Rev.
77
,
451
460
(
2017
).
58.
J.
Mandal
 et al., “
Hierarchically porous polymer coatings for highly efficient passive daytime radiative cooling
,”
Science
362
,
315
319
(
2018
).
59.
T. L.
Bergman
,
F. P.
Incropera
,
D. P.
DeWitt
, and
A. S.
Lavine
,
Fundamentals of Heat and Mass Transfer
(
John Wiley & Sons
,
2011
).
60.
G.
Oliveti
,
N.
Arcuri
,
M.
De Simone
, and
R.
Bruno
, “
Experimental evaluations of the building shell radiant exchange in clear sky conditions
,”
Sol. Energy
86
,
1785
1795
(
2012
).
61.
T. M.
Crawford
and
C. E.
Duchon
, “
An improved parameterization for estimating effective atmospheric emissivity for use in calculating daytime downwelling longwave radiation
,”
J. Appl. Meteorol.
38
,
474
480
(
1999
).
62.
J. A.
Voogt
and
T. R.
Oke
, “
Thermal remote sensing of urban climates
,”
Remote Sens. Environ.
86
,
370
384
(
2003
).
63.
N. G.
Loeb
 et al., “
Observed changes in top-of-the-atmosphere radiation and upper-ocean heating consistent within uncertainty
,”
Nat. Geosci.
5
,
110
113
(
2012
).
64.
A. K.
AAngström
,
A Study of the Radiation of the Atmosphere: Based upon Observations of the Nocturnal Radiation during Expeditions to Algeria and to California
(
Smithsonian Institution
,
1915
), Vol.
65
.
65.
D.
Brunt
, “
Notes on radiation in the atmosphere. I
,”
Q. J. R. Meteorol. Soc.
58
,
389
420
(
1932
).
66.
H. E.
Bennett
,
J. M.
Bennett
, and
M. R.
Nagel
, “
Distribution of infrared radiance over a clear sky
,”
JOSA
50
,
100
106
(
1960
).
67.
R. M.
Goody
,
Atmospheric Radiation Volume 1: Theoretical Basis
(
Clarendon Press
,
Oxford
,
1964
).
68.
R. W.
Bliss
, Jr.
, “
Atmospheric radiation near the surface of the ground: A summary for engineers
,”
Sol. Energy
5
,
103
120
(
1961
).
69.
W. C.
Swinbank
, “
Long-wave radiation from clear skies
,”
Q. J. R. Meteorol. Soc.
89
,
339
348
(
1963
).
70.
S. B.
Idso
and
R. D.
Jackson
, “
Thermal radiation from the atmosphere
,”
J. Geophys. Res.
74
,
5397
5403
, (
1969
).
71.
W.
Brutsaert
, “
On a derivable formula for long-wave radiation from clear skies
,”
Water Resour. Res.
11
,
742
744
, (
1975
).
72.
Cold Facts on Global Warming
, http://www.randombio.com/co2.html , see Figure 4 for absorption of ultraviolet, visible, and infrared radiation by various gases in the atmosphere; accessed 7 August
2018
.
73.
J. P.
Peixoto
and
A. H.
Oort
,
Physics of Climate
(
AIP-Press
,
1992
).
74.
E. M.
Lushiku
,
A.
Hjortsberg
, and
C. G.
Granqvist
, “
Radiative cooling with selectively infrared-emitting ammonia gas
,”
J. Appl. Phys.
53
,
5526
5530
(
1982
).
75.
J. T.
Kiehl
and
K. E.
Trenberth
, “
Earth's annual global mean energy budget
,”
Bull. Am. Meteorol. Soc.
78
,
197
208
(
1997
).
76.
C.
Riordan
and
R.
Hulstron
, “
What is an air mass 1.5 spectrum? (solar cell performance calculations)
,” in
IEEE Conference on Photovoltaic Specialists
(
1990
), Vol.
2
, pp.
1085
1088
.
77.
Solar Spectral Irradiance
: Air Mass 1.5, https://rredc.nrel.gov/solar//spectra/am1.5/#about for Air mass (AM) 1.5 direct solar spectrum; accessed 19 February
2019
.
78.
E.
Rephaeli
,
A.
Raman
, and
S.
Fan
, “
Ultrabroadband photonic structures to achieve high-performance daytime radiative cooling
,”
Nano Lett.
13
,
1457
1461
(
2013
).
79.
X.
Sun
,
Y.
Sun
,
Z.
Zhou
,
M. A.
Alam
, and
P.
Bermel
, “
Radiative sky cooling: Fundamental physics, materials, structures, and applications
,”
Nanophotonics
6
,
997
(
2017
).
80.
C. N.
Awanou
, “
Clear sky emissivity as a function of the zenith direction
,”
Renewable Energy
13
,
227
248
(
1998
).
81.
X.
Berger
and
J.
Bathiebo
, “
Directional spectral emissivities of clear skies
,”
Renewable Energy
28
,
1925
1933
(
2003
).
82.
A. W.
Harrison
, “
Effect of atmospheric humidity on radiation cooling
,”
Sol. Energy
26
,
243
247
(
1981
).
83.
R. A.
Oetjen
,
E. E.
Bell
,
J.
Young
, and
L.
Eisner
, “
Spectral radiance of sky and terrain at wavelengths between 1 and 20 microns. I. Instrumentation
,”
JOSA
50
,
1308
1313
(
1960
).
84.
T. J.
Garrett
,
L. F.
Radke
, and
P. V.
Hobbs
, “
Aerosol effects on cloud emissivity and surface longwave heating in the arctic
,”
J. Atmos. Sci.
59
,
769
778
(
2002
).
85.
X.
Berger
,
J.
Bathiebo
,
F.
Kieno
, and
C. N.
Awanou
, “
Clear sky radiation as a function of altitude
,”
Renewable Energy
2
,
139
157
(
1992
).
86.
X.
Berger
,
B.
Cubizolles
, and
I.
Donet
, “
Radio-sounding data for the determination of the infra-red sky radiation
,”
Sol. Wind Technol.
5
,
353
363
(
1988
).
87.
L.
Alados-Arboledas
and
J. I.
Jimenez
, “
Day-night differences in the effective emissivity from clear skies
,”
Boundary-Layer Meteorol.
45
,
93
101
(
1988
).
88.
Y.
Iijima
and
M.
Shinoda
, “
The influence of seasonally varying atmospheric characteristics on the intensity of nocturnal cooling in a high mountain hollow
,”
J. Appl. Meteorol.
41
,
734
743
(
2002
).
89.
C.
Ruckstuhl
,
R.
Philipona
,
J.
Morland
, and
A.
Ohmura
, “
Observed relationship between surface specific humidity, integrated water vapor, and longwave downward radiation at different altitudes
,”
J. Geophys. Res. Atmos.
112
,
D03302
, (
2007
).
90.
W. B.
Grant
, “
Water vapor absorption coefficients in the 8–13-μm spectral region: A critical review
,”
Appl. Opt.
29
,
451
462
(
1990
).
91.
J.
Duan
 et al., “
GPS meteorology: Direct estimation of the absolute value of precipitable water
,”
J. Appl. Meteorol.
35
,
830
838
(
1996
).
92.
A. W.
Harrison
and
M. R.
Walton
, “
Radiative cooling of TiO2 white paint
,”
Sol. Energy
20
,
185
188
(
1978
).
93.
C. Y.
Tso
,
K. C.
Chan
, and
C. Y. H.
Chao
, “
A field investigation of passive radiative cooling under Hong Kong's climate
,”
Renewable Energy
106
,
52
61
(
2017
).
94.
T.
Suichi
,
A.
Ishikawa
,
Y.
Hayashi
, and
K.
Tsuruta
, “
Performance limit of daytime radiative cooling in warm humid environment
,”
AIP Adv.
8
,
055124
(
2018
).
95.
P.
Berdahl
and
M.
Martin
, “
Emissivity of clear skies
,”
Sol. Energy
32
,
663
664
(
1984
).
96.
D. W.
Deering
and
P.
Leone
, “
A sphere-scanning radiometer for rapid directional measurements of sky and ground radiance
,”
Remote Sens. Environ.
19
,
1
24
(
1986
).
97.
E. M.
Lushiku
and
C.-G.
Granqvist
, “
Radiative cooling with selectively infrared-emitting gases
,”
Appl. Opt.
23
,
1835
1843
(
1984
).
98.
A.
Skartveit
,
J. A.
Olseth
,
G.
Czeplak
, and
M.
Rommel
, “
On the estimation of atmospheric radiation from surface meteorological data
,”
Sol. Energy
56
,
349
359
(
1996
).
99.
J. E. A.
Selby
and
R. M.
McClatchey
,
Atmospheric Transmittance from 0.25 to 28.5/micrometers Computer Code LOWTRAN 2
(
Air Force Cambridge Research Labs L G
,
Hanscom Field, MA
,
1972
).
100.
J. E.
Selby
and
R. A.
McClatchey
,
Atmospheric Transmittance from 0.25 to 28.5 Microns: Computer Code LOWTRAN 3
(
Air Force Cambridge Research Labs
,
Hanscom AFB MA
,
1975
).
101.
O. P.
Manley
,
H. J. P.
Smith
,
Y. M.
Treve
,
J. W.
Carpenter
, and
T. C.
Degges
,
OPTIR II
(
Visidyne Inc
.,
Woburn, MA
,
1971
).
102.
J. E. A.
Selby
,
F. X.
Kneizys
,
J. H.
Chetwynd
, and
R. A.
McClatchey
,
Atmospheric Transmittance/Radiance: Computer Code LOWTRAN 4
(
Air Force Geophysics Lab
,
Hanscom AFB, MA
,
1978
).
103.
T. S.
Eriksson
and
C. G.
Granqvist
, “
Radiative cooling computed for model atmospheres
,”
Appl. Opt.
21
,
4381
(
1982
).
104.
F. X.
Kneizys
,
E. P.
Shettle
, and
W. O.
Gallery
, “
Atmospheric transmittance and radiance: The LOWTRAN 5 code
,” in
Atmospheric Transmission
(
International Society for Optics and Photonics
,
1981
), Vol.
0277
, pp.
116
125
.
105.
X.
Berger
and
J.
Bathiebo
, “
From spectral clear sky emissivity to total clear sky emissivity
,”
Sol. Wind Technol.
6
,
551
556
(
1989
).
106.
A. J.
Prata
, “
A new long-wave formula for estimating downward clear-sky radiation at the surface
,”
Q. J. R. Meteorol. Soc.
122
,
1127
1151
(
1996
).
107.
F. X.
Kneizys
,
E. P.
Shettle
,
L. W.
Abreu
,
J. H.
Chetwynd
, and
G. P.
Anderson
,
Users Guide to LOWTRAN 7
(
Air Force Geophysics Lab
,
Hanscom AFB, MA
,
1988
).
108.
A.
Berk
,
L. S.
Bernstein
, and
D. C.
Robertson
,
MODTRAN: A Moderate Resolution Model for LOWTRAN
(
Spectral Sciences Inc.
,
Burlington, MA
,
1987
).
109.
MODTRAN®
, http://modtran.spectral.com/ for introduction about MODTRAN; accessed 30 September
2018
.
110.
S. D.
Lord
, “
A new software tool for computing Earth's atmospheric transmission of near- and far-infrared radiation
,”
Report No. NASA-TM-103957
(
1992
).
111.
S. A.
Lord
, https://atran.sofia.usra.edu/cgi-bin/atran/atran.cgi for introduction about ATRAN; accessed 4 April 2018 (
1992
).
112.
D. O.
Staley
and
G. M.
Jurica
, “
Effective atmospheric emissivity under clear skies
,”
J. Appl. Meteorol.
11
,
349
356
(
1972
).
113.
M.
Centeno
, “
New formulae for the equivalent night sky emissivity
,”
Sol. Energy
28
,
489
498
(
1982
).
114.
R.
Tang
,
Y.
Etzion
, and
I. A.
Meir
, “
Estimates of clear night sky emissivity in the Negev Highlands, Israel
,”
Energy Convers. Manage.
45
,
1831
1843
(
2004
).
115.
P.
Berdahl
,
M.
Martin
, and
F.
Sakkal
, “
Thermal performance of radiative cooling panels
,”
Int. J. Heat Mass Transfer
26
,
871
880
(
1983
).
116.
M.
Li
,
Y.
Jiang
, and
C. F. M.
Coimbra
, “
On the determination of atmospheric longwave irradiance under all-sky conditions
,”
Sol. Energy
144
,
40
48
(
2017
).
117.
S. B.
Idso
, “
A set of equations for full spectrum and 8- to 14-μm and 10.5- to 12.5-μm thermal radiation from cloudless skies
,”
Water Resour. Res.
17
,
295
304
, (
1981
).
118.
S.
Vall
and
A.
Castell
, “
Radiative cooling as low-grade energy source: A literature review
,”
Renewable Sustainable Energy Rev.
77
,
803
820
(
2017
).
119.
M.
Zeyghami
,
D. Y.
Goswami
, and
E.
Stefanakos
, “
A review of clear sky radiative cooling developments and applications in renewable power systems and passive building cooling
,”
Sol. Energy Mater. Sol. Cells
178
,
115
128
(
2018
).
120.
B.
Zhao
,
M.
Hu
,
X.
Ao
,
N.
Chen
, and
G.
Pei
, “
Radiative cooling: A review of fundamentals, materials, applications, and prospects
,”
Appl. Energy
236
,
489
513
(
2019
).
121.
L.
Evangelisti
,
C.
Guattari
, and
F.
Asdrubali
, “
On the sky temperature models and their influence on buildings energy performance: A critical review
,”
Energy Build.
183
,
607
625
(
2019
).
122.
R.
Vidhi
, “
A review of underground soil and night sky as passive heat sink: Design configurations and models
,”
Energies
11
,
2941
(
2018
).
123.
E. L.
Andreas
and
S. F.
Ackley
, “
On the differences in ablation seasons of arctic and Antarctic Sea ice
,”
J. Atmos. Sci.
39
,
440
447
(
1982
).
124.
X.
Berger
,
D.
Buriot
, and
F.
Garnier
, “
About the equivalent radiative temperature for clear skies
,”
Sol. Energy
32
,
725
733
(
1984
).
125.
A. C.
Dilley
and
D. M.
O'brien
, “
Estimating downward clear sky long-wave irradiance at the surface from screen temperature and precipitable water
,”
Q. J. R. Meteorol. Soc.
124
,
1391
1401
(
1998
).
126.
S.
Niemelä
,
P.
Räisänen
, and
H.
Savijärvi
, “
Comparison of surface radiative flux parameterizations: Part I: Longwave radiation
,”
Atmos. Res.
58
,
1
18
(
2001
).
127.
J. P.
Lhomme
,
J. J.
Vacher
, and
A.
Rocheteau
, “
Estimating downward long-wave radiation on the Andean Altiplano
,”
Agric. Meteorol.
145
,
139
148
(
2007
).
128.
J. E.
Sicart
,
R.
Hock
,
P.
Ribstein
, and
J. P.
Chazarin
, “
Sky longwave radiation on tropical Andean glaciers: Parameterization and sensitivity to atmospheric variables
,”
J. Glaciol.
56
,
854
860
(
2010
).
129.
B. A.
Kimball
,
S. B.
Idso
, and
J. K.
Aase
, “
A model of thermal radiation from partly cloudy and overcast skies
,”
Water Resour. Res.
18
,
931
936
, (
1982
).
130.
F.
Kasten
and
G.
Czeplak
, “
Solar and terrestrial radiation dependent on the amount and type of cloud
,”
Sol. Energy
24
,
177
189
(
1980
).
131.
J.
Schmetz
, “
Towards a surface radiation climatology: Retrieval of downward irradiances from satellites
,”
Atmos. Res.
23
,
287
321
(
1989
).
132.
P.
Schmetz
,
J.
Schmetz
, and
E.
Raschke
, “
Estimation of daytime downward longwave radiation at the surface from satellite and grid point data
,”
Theor. Appl. Climatol.
37
,
136
149
(
1986
).
133.
T.
Saitoh
and
T.
Ono
, “
Utilization of seasonal sky radiation energy for space cooling
,”
J. Sol. Energy Eng.
106
,
403
407
(
1984
).
134.
M.
Sugita
and
W.
Brutsaert
, “
Cloud effect in the estimation of instantaneous downward longwave radiation
,”
Water Resour. Res.
29
,
599
605
, (
1993
).
135.
P.
Chylek
and
V.
Ramaswamy
, “
Simple approximation for infrared emissivity of water clouds
,”
J. Atmos. Sci.
39
,
171
177
(
1982
).
136.
P. J.
Robinson
and
J. A.
Davies
, “
Laboratory determinations of water surface emissivity
,”
J. Appl. Meteorol.
11
,
1391
1393
(
1972
).
137.
M.
Aubinet
, “
Longwave sky radiation parametrizations
,”
Sol. Energy
53
,
147
154
(
1994
).